Podcasts about paragraphs

OLG Nürnberg 14.04.2026 - 11 UF 940/25: Abonnieren und weiter empfehlen!Instagram: rechtsprechung_newsJura; Urteil; Rechtsprechung; News; Referendariat;Rechtswissenschaften; Prozess; Recht; Gericht; Gesetz; Klage;Rechtsanwalt; Staatsexamen; Paragraf; Jurist; Examen; StEx;Rechtsreferendariat; Anwalt; Ref; Paragraph; Referendar; Justiz; Bundesverfassungsgericht; Rechtsreferendar; Richter; law; Jurastudent; Jurapodcast; Staatsanwalt; Rechtswissenschaft; Streit; Verurteilung; Polizei; Beamte; Polizist; Klage; Kläger; Beklagte; Klausur; Erstesexamen; Assessorexamen; Erstesstaatsexamen; Repetitor; Repetitorium; Assessor; Zivilrecht; BGB; BGH; Bundesgerichthof; Landgericht; Oberlandesgericht; OLG; LG; Amtsgericht; AG; ZPO; Strafrecht; StGB; Strafgesetzbuch; Strafe; StPO; strafbar; Bewährung; Beschlagnahme; Prozess; Haftung; Rechtsstaat; Demokratie; zivil; Verbraucher; Unternehmer; Widerruf; Unternehmen; Grundrechte;Strafverteidiger; Grundgesetz; Strand; Hochzeit; Geschenk; Hochzeitsgeschenk; Ehe; Familie; Familienrecht; Sachenrecht; Mobiliar; KfZ; PKW; Fahrzeug; Auto; Schenkung; Fahrzeugbrief; Zulassung; Bescheinigung; Übereignung; Familiengericht; Trennung; Scheidung; Herausgabe;

//The Wire//2300Z June 17, 2026// //ROUTINE// //BLUF: UNITED STATES RELEASES TERMS OF MOU TO END THE GULF WAR. CONFUSION ABOUNDS REGARDING STATUS OF MERCHANT SHIPPING IN THE STRAIT OF HORMUZ. INTERSTATE SHOOTING SPREE CONDUCTED IN KANSAS CITY. RAPE GANG INQUIRY REPORT RELEASED IN UNITED KINGDOM.// -----BEGIN TEARLINE------International Events-Middle East: This morning, the United States published the text of the upcoming Memorandum of Understanding, which is to be signed on Friday. This disclosure was made by American officials to CNN and Bloomberg at the G7 summit, who verbally read the terms and provided the text to MSM outlets. Per this text, the general gist of the deal is very similar to what the Iranians claim, but with some minor differences. The most contentious item of the deal is right up front in Paragraph 1: the war in Lebanon is included as part of the deal. All sanctions on Iran will also be lifted, and a $300 billion reparations payment will be funded by the United States and other Arab nations for the reconstruction of Iran's infrastructure that was destroyed in the war. President Trump also verbally stated that Iran's missile program will remain, and is not part of the deal.Strait of Hormuz: The status of shipping remains uncertain as both sides continue to enforce the mutual blockades. Several Iranian ships have crossed the American blockade line, despite NAVCENT stating that the blockade is still in effect, but western-aligned ships have been hesitant to make the crossing due to the confusion.Analyst Comment: Right now a lot of commercial firms and insurance companies are trying to figure out what to do, and a surge of petroleum tankers is heading toward the Middle East right now, as the world awaits the resumption of normal shipping this weekend. The US and Iran have thirty days after Friday to allow shipping to resume, but 60 days after the agreement is signed, Iran will retain the right to charge tolls for access to the Persian Gulf, granting the Iranians de facto control of the waterway. As a result, the insurance status of commercial shipping remains unclear, though rates will very likely be permanently elevated, much like how many shipping companies never reverted back to the Red Sea route following the Houthi targeting two years ago. Many companies still take the longer route around Africa, so that situation has not yet returned to normal, and though alternative access routes for the Persian Gulf don't really exist, it will take many months for shipping to work out the details of how to proceed.-HomeFront-Missouri: Last night, a mass shooting spree was reported throughout Kansas City as one assailant conducted small arms attacks at five different sites throughout the city. Local authorities state that all of these shootings stem from the same incident, which appears to involve an individual in a vehicle, traveling eastbound on I-70 shooting at other vehicles also traveling the same direction. Later that night, a fifth shooting site was located after a man was found wounded in his vehicle at the intersection of Truman Road and Bennington Avenue. This man later died at a local hospital.All total, five different vehicles were targeted, and a total of four people were wounded, and one person killed, during the attacks throughout the day. By this morning, the link between all shootings was discovered, and the suspect was located at his residence in Independence, where a barricaded-shooter situation is currently underway. More details are expected as the situation develops.-----END TEARLINE-----Analyst Comments: In the United Kingdom, the Rape Gang Inquiry Report was released last night. This report was compiled by a Parliamentary effort led by MP Rupert Lowe, with the goal of investigating the official documents and criminal cases of mass the rape of British society by Pakistani men since the 1950's. Specifically, the investigation seeks to examine the industrialization of rape, organized into cells or Rape Gangs, which conducted assaults on a scale never before seen by the civilized world.This report revealed that an minimum of 250,000 British girls have been systemically raped, tortured, and victimized by these gangs, which included a large percentage of police officers, judges, and Members of Parliament. An entire ecosystem had been established for the police to arrest any victims of the gangs, with victims in most cases being raped hundreds of times by police officers while allegedly being in victim protection programs, even after the scandal became public years ago. Nearly 40% of the land mass of the nation was host to these gangs, which were not conducting isolated attacks, but industrial-scale crimes at a volume that is simply indescribable.The 200+ page report is not for the faint of heart as it describes in excruciating detail the evidence examined in this case, and nearly all of it is too horrific for mixed company. The closest approximation for a wider audience would be to imagine a network of hundreds of thousands of Jeffrey Epsteins, committing acts that were so horrific that Epstein's own network did not engage in this level of depravity. For most people, this document will be the absolute worst thing they will ever read in their entire life.Right now, the western world is at an impasse. The recent social tension in the U.K. due to the stabbing attacks has served as a primer for whatever comes next. Understanding the severity of this report, the National Crime Agency has snapped-to, and ordered the re-investigation of rape gang reports going back to 2010, in an attempt to get ahead of the tidal wave of righteous anger that has begun to rise throughout society. The entire Child Services ecosystem of the United Kingdom is currently serving as a clearing house to funnel a quarter of a million children to Muslim rape gangs. Kier Starmer himself was the director of Public Prosecutions during the height of this atrocity, and personally signed off on ~13,000 rape gang members being released with nothing more than a warning letter.If these people were sent a letter, that means they have names and addresses. It will be for the British people to decide how best to proceed, and it will take some time for organizational efforts to take hold, considering the sheer scale of this crisis. It will also be wise to consider that this problemset is not unique (nor contained) to the British Isles.Analyst: S2A1 Research: https://publish.obsidian.md/s2underground Disclaimer: No LLMs were used in the writing of this report. //END REPORT//

Mark Dubowitz, CEO of the Foundation for Defense of Democracies, returns to School of War to discuss the newly released memorandum of understanding between the United States and Iran. What does the deal actually say? What are its biggest strategic implications? And is it comparable to the JCPOA—or something worse? 01:19 - Paragraph 1: Lebanon and Hezbollah 09:45 - Paragraphs 2–3: Sovereignty and the 60-day clock 11:24 - Paragraph 4: Lifting the blockade 16:37 - Paragraph 5: The Strait of Hormuz 23:48 - Paragraph 6: A $300 billion reconstruction fund 33:49 - Paragraph 7: Ending all sanctions 35:43 - Paragraph 8: Nuclear weapons and enrichment 39:12 - Paragraph 9: Maintaining the status quo 40:44 - Paragraph 10: Oil waivers and sanctions relief 43:48 - Paragraph 11: Releasing frozen funds 46:56 - Paragraph 12: Monitoring compliance 47:52 - Paragraph 13: Negotiating the final deal 50:59 - Paragraph 14: A UN Security Council resolution 51:13 - Squandering battlefield leverage Follow along on Instagram, X @schoolofwarpod, and YouTube @SchoolofWarPodcast Find more at The Free Press. Learn more about your ad choices. Visit megaphone.fm/adchoices

Pastor David continues our series in the Sermon on the Mount with his exposition of Matthew 5:7.(7A) THE CONDITION AND ATTITUDE OF THE MERCIFUL;Colossians 1:13, 2:6-7a,Titus 3:3-6,Exodus 34:6,1 Peter 1:3,Romans 5:8,Micah 6:8 (NKJV),Colossians 3:12,Luke 6:36,Ephesians 5:1,James 3:17,Jude 1:20-23 (CSB),2 Timothy 1:16,Proverbs 14:21,Luke 10:3-37 (NIV),Second London Baptist Confession,Chapter 22, Paragraph 8(V7B) (AND) THE PROMISED OUTCOME OF THE MERCIFUL.Psalm 18:25,Hebrews 4:16,2 Timothy 1:18,Jude 1:1b-2Second London Baptist Confession,Chapter 32, Paragraph 2

Sermon Outline: "On My Father's Side" Preacher: Pastor Lemuel Miller (Guest Speaker / Advisory Board Member) Location: Church of the Harvest I. Introduction: The Temple and the Root Causes of Sickness The Caleb Spirit: At nearly 74 years old, Pastor Lemuel shares his experience winning a silver medal at the national arm wrestling championship, emphasizing that our bodies are the temple of the Holy Spirit (1 Corinthians 6:19). We must actively care for our physical templates so that the Holy Ghost is not "living in a garage or a shed." Commanding Prayers vs. Begging: In his book Prayers for Healing, Pastor Lemuel explains that many believers stay in ignorance, begging God for things He has already accomplished. Spiritual maturity requires switching from begging prayers to authoritative, commanding prayers. Uprooting Sickness: True physical restoration requires looking past the surface pain and identifying the structural root causes of diseases. Migraine Headaches: Often rooted in trauma and bitterness. Kidney Disease: Rooted in deep unforgiveness, bitterness, fear, and self-rejection. Leukemia: Often linked to bitterness and paternal rejection. Lupus: Can stem from deep-seated guilt, self-hatred, and low self-esteem. II. Point 1: Understanding Jesus as a 100% Human Example The Human Dependency: Reading from John 5:19, Jesus explicitly declares that the Son can do absolutely nothing of Himself except what He sees the Father do. Many Christians incorrectly attribute Jesus' earthly miracles to His inherent divinity, forgetting that He stripped Himself of that privilege to come as a 100% vulnerable human baby. The Earthly Blueprint: Jesus had to fully depend on human care and look directly to His Heavenly Father for supernatural strength. By doing this, He serves as a complete human blueprint for how we are meant to walk out authority on Earth. III. Point 2: The Full Meaning of Salvation (Sozo) Fire Insurance vs. Full Rights: Most modern believers view being "saved" purely as a post-death ticket to Heaven to avoid Hell. The Greek Meaning: In Matthew 1:21, the term for save is the Greek word Sozo. When properly translated, it means you are actively rescued from: Accidents, injuries, physical harm, and structural danger. Destruction, risk, peril, loss, and premature/untimely death. Sickness, chronic disease, physical infirmity, and generational curses. The Transacted Benefits: True Sozo simultaneously grants the believer legal rights to divine prosperity, total deliverance from addictions, inner strength, structural healing, and operational wholeness. IV. Point 3: The Conversation in the Temple (The Two Sides) Using a vivid exploration of Luke 2:47, Pastor Lemuel illustrates the 12-year-old Jesus sitting among the elite rabbis and theologians, contrasting His dual lineage: On My Mother's Side: Born into natural law, generational trauma, human limitations, fear, hunger, thirst, trouble, and the lingering curse of sin. On My Father's Side: Formed in supernatural law, eternal life, and generational blessings. The Supernatural Exchange: * Hungry/Thirsty: On His mother's side, He fasts; on His Father's side, He is the Bread of Life and a well that never runs dry. Natural Law: His mother's side is bound to gravity; His Father sits upon the flood (Psalm 29:10)—and a Father who sits on the flood raises a Son who walks on water. Surrounding Defense: On His Father's side, He is covered by feathers and wings (Psalm 91:4), meaning His truth functions as a structural shield, buckler, and surrounding defense against the snare of the fowler. Age and Identity: On His mother's side, He is a 12-year-old from Nazareth; on His Father's side, He is the Ancient of Days, Alpha and Omega, the Architect of the Universe, and the Lamb slain before the foundation of the world. V. Point 4: Operating in Dunamis Power The Age of Public Service: In Hebrew culture, a priest could not step into public high-priestly service until age 30. Accordingly, Jesus did not perform public miracles until reaching this baseline. The Necessity of Anointing: Acts 10:38 states that God anointed Jesus of Nazareth with the Holy Ghost and power. If Jesus was operating strictly as God, He wouldn't require an anointing or a companion. He operated as an anointed human being. Miraculous Power (Dunamis): When the Holy Ghost fills a believer, they receive Dunamis power—the explosive, dynamic, and supernatural capacity to perform miracles. Rebuking the Root: When dealing with demonic possession (Acts 16), Peter's mother-in-law's fever (Luke 4), or the raging sea (Mark 4), Jesus always used sharp, severe, and authoritative rebukes (epitimao). In the storm, He did not rebuke the water; He rebuked the wind—the structural root cause of the problem. VI. Conclusion: Activating Faith vs. Waiting The Whip Post Transaction: Based on Isaiah 53:5 and 1 Paragraph 2:24, our healing is already completely provided in the past tense ("by whose stripes you were healed"). Healing was legally settled at the whipping post, not two years from now. Faith is Practical Action: Believers fail to receive because they allow passive doubt to eat their seed of faith. Like the ten lepers in Luke 17, their structural healing manifested on the way because they actively moved in obedience to Jesus' command. Healing vs. Wholeness: While nine lepers were cleansed (cured of the disease), the Samaritan leper who turned back to worship Jesus was made whole (Sozo). Healing cures the virus; wholeness creatively restores every limb, finger, or piece of flesh that was eaten away or missing. Scripture Index Here are the primary scriptures read, cited, or expounded upon during the service: Deuteronomy 34:7 (Referenced) – The account of Moses being 120 years old with eyes undimmed and his natural force unabated. 1 Corinthians 6:19-20 (Referenced) – Knowing that your body is the temple of the Holy Ghost, bought with a price. John 5:19 > "Then answered Jesus and said unto them, Verily, verily, I say unto you, The Son can do nothing of himself, but what he seeth the Father do: for what things soever he doeth, these also doeth the Son likewise." Isaiah 7:14 > "Therefore the Lord himself shall give you a sign; Behold, a virgin shall conceive, and bear a son, and shall call his name Immanuel." Isaiah 9:6 > "For unto us a child is born, unto us a son is given: and the government shall be upon his shoulder: and his name shall be called Wonderful, Counsellor, The mighty God, The everlasting Father, The Prince of Peace." Matthew 1:21 > "And she shall bring forth a son, and thou shalt call his name Jesus: for he shall save [sozo] his people from their sins." Luke 2:8-14 (Paraphrased) – The angelic announcement to the shepherds keeping watch over their flocks by night in the city of David. Luke 2:47 > "And all that heard him were astonished at his understanding and answers." Romans 8:2 (Referenced) – The law of the Spirit of life in Christ Jesus making us free from the law of sin and death. Psalm 29:10 (Referenced) – The Lord sitting upon the flood; the Lord sitting King forever. Psalm 91:1-16 (Completely Quoted) – The structural promises of protection, including abiding under the shadow of the Almighty, delivery from the snare of the fowler, protection from night terror, and angels bearing the believer up. Psalm 8:4-8 (Referenced) – What is man that thou art mindful of him, creating him a little lower than the angels and putting all things under his feet. Luke 10:19 > "Behold, I give unto you power to tread on serpents and scorpions, and over all the power of the enemy: and nothing shall by any means hurt you." Psalm 103:1-5 (Referenced) – Blessing the Lord and forgetting not His benefits, who forgives iniquities and heals all diseases. Mark 11:23 > "For verily I say unto you, That whosoever shall say unto this mountain, Be thou removed, and be thou cast into the sea; and shall not doubt in his heart... he shall have whatsoever he saith." Acts 10:38 > "How God anointed Jesus of Nazareth with the Holy Ghost and with power: who went about doing good, and healing all that were oppressed of the devil; for God was with him." Isaiah 53:5 > "But he was wounded for our transgressions, he was bruised for our iniquities: the chastisement of our peace was upon him; and with his stripes we are healed." 1 Peter 2:24 > "Who his own self bare our sins in his own body on the tree, that we, being dead to sins, should live unto righteousness: by whose stripes ye were healed." Hebrews 1:14 > "Are they not all ministering spirits, sent forth to minister for them who shall be heirs of salvation?" 2 Corinthians 5:21 > "For he hath made him to be sin for us, who knew no sin; that we might be made the righteousness of God in him." Luke 17:11-19 (Referenced) – The healing of the ten lepers on their way to the priest, and the structural wholeness given to the one returning Samaritan. "Thanks for listening! For more information, visit churchoftheharvest.com. Don't forget to follow us on Facebook and YouTube @cothcleveland.

Summary Coming: page 44b-45a; Paragraph 11-13Beginning of R' Nosson

2 LBCF Chapter 5, Paragraphs 1-3

Save up $100 on our B2 English Fluency course. Offer expires April 7th. Go here to enroll today. Learn more about your ad choices. Visit podcastchoices.com/adchoices

SUBSCRIBE to The FoundHer Files, our twice weekly Substack filled with actionable tips you can use starting today to build and grow your business. No fluff. No gatekeeping. Just what works. Host, Lindsay Pinchuk provides a tactical walkthrough on crafting a compelling three-paragraph founder story that effectively communicates your business, builds trust, and attracts the right audience. This episode offers a step-by-step framework, common mistakes to avoid, and practical tips for using your story across various platforms.Grab this week's Deep Dive on The FoundHer Files to read more about The Anatomy of a Story that Converts. Join us for this month's Forum Expert Workshop: Claude for FoundHers with Dara Astmann REGISTER HERESubscribe to The FoundHer Files Substack: http://foundherfiles.substack.comJoin our online networking community for women business owners over forty, The Dear FoundHer... Forum.Follow Dear FoundHer... on Instagram http://www.instagram.com/dearfoundher Hosted on Acast. See acast.com/privacy for more information.

BGH 24.02.2026 – VI ZR 415/23 und VI ZR 416/23:Abonnieren und weiter empfehlen! Instagram: rechtsprechung_newsJura; Urteil; Rechtsprechung; News; Referendariat;Rechtswissenschaften; Prozess; Recht; Gericht; Gesetz; Klage;Rechtsanwalt; Staatsexamen; Paragraf; Jurist; Examen; StEx;Rechtsreferendariat; Anwalt; Ref; Paragraph; Referendar; Justiz; Bundesverfassungsgericht; Rechtsreferendar; Richter; law; Jurastudent; Jurapodcast; Staatsanwalt; Rechtswissenschaft; Streit; Verurteilung; Polizei; Beamte; Polizist; Klage; Kläger; Beklagte; Klausur; Erstesexamen; Assessorexamen; Erstesstaatsexamen; Repetitor;Repetitorium; Assessor; Zivilrecht; BGB; BGH; Bundesgerichthof; Landgericht; Oberlandesgericht; OLG; LG; Amtsgericht; AG; ZPO; Strafrecht; StGB; Strafgesetzbuch; Strafe; StPO; strafbar; Bewährung; Beschlagnahme; Prozess; Haftung; Rechtsstaat; Demokratie; zivil; Verbraucher; Unternehmer; Widerruf; Unternehmen; Grundrechte;Strafverteidiger; Grundgesetz; Marokko; Staat; Unterlassung: Ehre; Spiegel; Zeit; Spionage; Pegasus; Portal; Verlag; Internet; Achtung; Völkerrecht; Königreich;

Wer Kanzler Friedrich Merz als „Lügenfritz“ bezeichnet, muss unter Umständen einen Monatslohn Strafe zahlen. So entschied es das Amtsgericht Öhringen in einem viel diskutierten Fall. Es ist nur das jüngste Beispiel in einer langen Reihe von Ermittlungen wegen Politikerbeleidigung. AUF1 wollte wissen, wie die Bürger darüber denken. Unser Reporter Roy Grassmann hat sich vor dem Kanzleramt in Berlin umgehört.

Steven Johnson dreamed of building the ultimate research assistant. Now he's doing just that at Google, where he's the co-founder and editorial director of NotebookLM.It's one of the most interesting AI products out there. It radically changes how we learn, research, and remember — and the "notebook" itself is becoming a standard unit of knowledge across Google, rolling out in more and more places where AI needs to reference a body of sources.In this episode, the author of _Where Good Ideas Come From_ explains how AI is making him a better researcher and writer — and why tools like NotebookLM are so powerful when you're trying to make new connections, remember what you've already found, and figure out what's missing.There's a lot of fear right now that AI is making us dumber. That by relying on it too much, we're engaging in "cognitive offloading" and stunting our learning. That's a real risk, especially in schools.But Steven says we should also be talking about what you can gain from AI — and the power of something he calls "cognitive uploading."Resources:* Google NotebookLM: https://notebooklm.google/* Steven Johnson: https://stevenberlinjohnson.com/Support Future Around & Find Out:* Follow Dan on LinkedIn https://www.linkedin.com/in/dblums/* Get the free newsletter: https://www.futurearound.com* Become a paid subscriber and help future proof FAFO! https://www.futurearound.com/upgrade(00:00) - If you are interested in truly understanding something, this is the greatest time to be alive (01:25) - Steven's controversial NYT piece and the cold call from Google Labs (02:55) - Who NotebookLM's power users are (04:40) - The notebook as a new format for knowledge (06:20) - Featured notebooks: earnings reports, Shakespeare, and Dungeons & Dragons (11:00) - Writing a book about the Gold Rush with NotebookLM (13:20) - Four weeks of research in 14 minutes (16:30) - Following serendipitous connections through the source material (17:50) - Cognitive offloading and the illusion of understanding (21:00) - How Steven actually writes with AI (24:30) - Paragraph by paragraph: a new kind of writing (26:55) - Do readers need to know AI helped write it? (28:55) - Where good ideas come from in the age of AI (31:56) - Searching the negative space (33:56) - The adjacent possible: custom software for everyone (37:01) - NotebookLM for nonprofits and small organizations (39:06) - Tens of thousands of quotes, 25 years of forgetting (40:56) - "It's cognitive uploading"

BGH 21.01.2026 – XII ZB 142/25: Abonnieren und weiter empfehlen! Instagram: rechtsprechung_newsJura; Urteil; Rechtsprechung; News; Referendariat;Rechtswissenschaften; Prozess; Recht; Gericht; Gesetz; Klage;Rechtsanwalt; Staatsexamen; Paragraf; Jurist; Examen; StEx;Rechtsreferendariat; Anwalt; Ref; Paragraph; Referendar; Justiz; Bundesverfassungsgericht; Rechtsreferendar; Richter; law; Jurastudent; Jurapodcast; Staatsanwalt; Rechtswissenschaft; Streit; Verurteilung; Polizei; Beamte; Polizist; Klage; Kläger; Beklagte; Klausur; Erstesexamen; Assessorexamen; Erstesstaatsexamen; Repetitor;Repetitorium; Assessor; Zivilrecht; BGB; BGH; Bundesgerichthof; Landgericht; Oberlandesgericht; OLG; LG; Amtsgericht; AG; ZPO; Strafrecht; StGB; Strafgesetzbuch; Strafe; StPO; strafbar; Bewährung; Beschlagnahme; Prozess; Haftung; Rechtsstaat; Demokratie; zivil; Verbraucher; Unternehmer; Widerruf; Unternehmen; Grundrechte;Strafverteidiger; Grundgesetz; Ehe; Scheidung; scheiden; trennen; Trennung; Unterhalt; Immobilie; Einfamilienhaus; Haus; Wohnung; vermieten; Miete; Vermieter; Nießbrauch; Kündigung; Zwangsversteigerung; versteigern; Familienrecht;

In this episode, we reflect on how Pentecost is more than just an event of the past, but can be an ongoing outpouring of God's love and grace in our lives today. We discuss how the Holy Spirit helps heal our wounds, restores communion with God, and teaches us how to live with power and boldness, even in the places where we feel unable to love the Lord. As we prepare our hearts for Pentecost, we talk practically about cultivating a posture of receptivity, learning to invite the Holy Spirit more deeply into our hearts, and how to allow His fire to illuminate our identity and transform us from within. The Holy Spirit is alive within us, able to enter even the locked places of our hearts, and desires to help us bear much fruit in our lives.   Heather's One Thing - Pentecost | Holy Spirit Rest on Us Playlist Heather's Other One Thing - Wild Goose Series with Fr. Dave Pivonka Sister Miriam's One Thing - The Discerning Hearts Podcast App Michelle's One Thing - College Graduations   Other Resources Mentioned: Be Transformed Book Study   Journal Questions: How can I make myself more open to encounter the Holy Spirit pirit this Pentecost? Where in my life do I feel powerless? Am I allowing the Holy Spirit to make me more like Jesus? How am I rationing the boldness of the Holy Spirit? Where do I need to unlock the doors of my heart?   Discussion Questions: How can you welcome the Holy Spirit into your life in a deeper way?  What does it look like to experience the Holy Spirit in community? What gifts of the Holy Spirit do you desire to empower your work of building the Kingdom of God? How will you celebrate Pentecost?   Quotes to Ponder: "If there be among the gifts of God none greater than love, and there is no greater gift of God than the Holy Spirit, what follows more naturally than that He is Himself love…." (Saint Augustine)   "This great mystical tradition . . . shows how prayer can progress, as a genuine dialogue of love, to the point of rendering the person wholly possessed by the divine Beloved, vibrating at the Spirit's touch, resting filially within the Father's heart." (Pope Saint John Paul II, Novo Millennio Ineunte, Paragraph 33)   "A noble and delicate soul ... follows faithfully the faintest breath of the Holy Spirit; it rejoices in this Spiritual Guest and holds onto Him like a child to its mother." (The Diary of St. Faustina, Entry 148) Scripture for Lectio: "On one occasion, while he was eating with them, he gave them this command: "Do not leave Jerusalem, but wait for the gift my Father promised, which you have heard me speak about. For John baptized with water, but in a few days you will be baptized with the Holy Spirit." Then they gathered around him and asked him, "Lord, are you at this time going to restore the kingdom to Israel?" He said to them: "It is not for you to know the times or dates the Father has set by his own authority. But you will receive power when the Holy Spirit comes on you; and you will be my witnesses in Jerusalem, and in all Judea and Samaria, and to the ends of the earth." After he said this, he was taken up before their very eyes, and a cloud hid him from their sight." (Acts 1:4-9)   Sponsor - Mary's Meals: Every mother knows what it's like to want to protect their child and provide for them, to make sure there's food on the table, to make sure her child is safe, cared for, and able to grow into who God created them to be. But for millions of mothers around the world, hunger stands in the way of that hope. Their children walk to school carrying empty bowls, wondering if today they'll have anything to eat. And that's where Mary's Meals steps in. Mary's Meals provides one daily meal in a place of education for children living in some of the world's poorest communities. And that simple meal becomes the reason a child comes to school. Once they're there, everything can begin to change. Education can become a pathway out of poverty.  So a child who is hungry can focus.  A child who is vulnerable can dream.  And a mother who felt helpless can begin to hope again.  What we love most about Mary's Meals is how ordinary people get to become a part of that story. We don't have to solve global hunger. We simply have to feed one child. And here's the beautiful thing. It only costs $25.20 to feed a child for an entire school year.  That's one child sitting in a classroom instead of sitting at home hungry. One mother experiencing relief instead of worry.  One life changed through a simple act of love. So if your hearts are moved to help, we invite you to join us in supporting Mary's Meals. You can head over to their website marysmealsusa.org (or marysmeals.ca for Canada) and together we can offer hope, dignity and a daily meal to a child who needs it most.    Timestamps: 00:00 Mary's Meals 01:39 Introduction 02:26 Welcome 03:24 Scripture Verse and Quote to Ponder 04:39 Preparing Our Hearts 07:34 Healing the Wound of Powerlessness  09:26 The Holy Spirit Enables Us to Love 11:04 Inviting the Holy Spirit into Our Hearts 12:50 Our Advocate and the Forgiveness of Sins 15:10 "It is Better that I Go" 17:14 Baptized with Fire 19:15 Bearing Fruit in Our Lives 20:57 When We Restrict the Holy Spirit 22:50 Receiving the Holy Spirit within Community 28:08 Being Inspired by the Holy Spirit 31:04 One Things

Fr. Eamonn McCarthy and Matthias Conroy continue their reading of Humanae Vitae, working through paragraphs 14 and 17. Paragraph 14 sets out the absolute exclusions: direct abortion, sterilisation, and any act specifically intended to prevent procreation. Paragraph 17 offers Paul VI's prophetic warnings about where the widespread acceptance of artificial contraception would lead — marital […] L'articolo Societal Damage of Contraception – Humanae Vitae – Church Wisdom with Fr Eamonn McCarthy & Matthias Conroy proviene da Radio Maria.

Worauf muss man sich im Sitzungsdienst noch einstellen? Na, auf Gespräche rund um die Einstellung! Warum und wann muss ich hier den Ausbilder anrufen, nach welchen Kriterien darf ich eine Einstellung mit und ohne Auflage für opportun halten - und wie behaupte ich mich, wenn ich zur Einstellung gedrängt werde? Herzstück dieser Folge ist aber etwas anderes: Das Plädoyer! Wie spreche ich die Anwesenden an? Wann plädiere ich eher lang, wann eher kurz, wann ist eine Beweiswürdigung geboten, wann eher nicht? Wie gehe ich mit meiner Nervosität um? Diese und viele weiteren Fragen erörterten Henning Barton, Richter am Landgericht, und Anna Henrichs, Richterin, beide AG-Leitende, in dieser dritten und letzten Folge unserer großen Reihe zum staatsanwaltschaftlichen Sitzungsdienst. Wenn ihr alle Folgen aufmerksam gehört habt, habt ihr neben vielen praktischen Tricke wie nebenbei auch noch etwas anderes gelernt: Nämlich mit welcher inneren und äußeren Haltung würdig und zugleich bestimmt unser Rechtsstaat vertreten werden kann. Denn auch das bedeutet Sitzungsvertretung - viel Spaß beim Hören! Kapitelmarken: 00:00 Begrüßung 04:40 Zeitpunkt der Einstellungsentscheidung 08:22 Einstellung oder Beschränkung? 13:20 Was mache ich, wenn ich zur Zustimmung „gedrängt“ werde? 16:30 „Vorprüfung“ bei Einstellung nach §§ 153, 153a StPO: Geständnis und Ersttäter 20:01 Argumente pro Einstellung 37:10 Ruf den Ausbilder an! 39:16 Die Beschränkungsmöglichkeiten nach §§ 154, 154a StPO 44:28 Das Plädoyer - kurz oder lang? 47:58 Der Aufbau des (langen) Plädoyers 50:10 Plädieren bitte! Ohne Pause? 58:11 Der erste Satz: Die Anrede 01:02:27 Die Sachverhaltsfeststellung 01:06:01 Beweise würdigen 01:06:55 Strafbar oder schuldig gemacht? 01:07:44 Die rechtliche Würdigung 01:08:37 Die Strafzumessung - ein Dreischritt 01:19:00 Basar - Was beantrage ich denn jetzt? 01:23:53 Vielen Dank! 01:24:48 Hennings Tipp gegen‘s Zittern 01:27:01 Das „kurze“ Plädoyer 01:35:29 Vorführung aus der U-Haft - was ist zu beachten? 01:42:00 Anträge nach Paragraph §§ 69, 69a StGB 01:48:52 Anträge auf Einziehung von Gegenständen 01:53:53 Und schon vorbei! 01:57:07 Verabschiedung http://www.refpod.de http://www.instagram.com/ref.pod/ E-Mail: jura.ref.pod@gmail.com Disclaimer: Der Podcast beinhaltet ausschließlich persönliche Ansichten der Podcasterinnen und Podcaster und insbesondere keine offiziellen Standpunkte der Justizprüfungsämter.

The Catechism begins Paragraph 5 and dives into the communion of saints. This communion is “in holy things (sancta)” and “among holy persons (sancti).” Fr. Mike emphasizes that “the least of our acts done in charity redounds to the profit of all.” At the same time, every sin hurts every single member of the Church. No one is an island—we belong to each other. Today's readings are Catechism paragraphs 946-953. This episode has been found to be in conformity with the Catechism by the Institute on the Catechism, under the Subcommittee on the Catechism, USCCB. For the complete reading plan, visit ascensionpress.com/ciy Please note: The Catechism of the Catholic Church contains adult themes that may not be suitable for children - parental discretion is advised.

Pastor David continues our series with his exposition of 1 Peter 5:8-11.(VV8-9) THE CALL TO BE WATCHFUL AND FIRMMark 14:47-38,Luke 22:31-34,1 Peter 1:13,1 Peter 4:7,Matthew 4:4,Psalm 119:9, 105,2 Timothy 3:16-17,Job 1:6-12,Psalm 10:8-9,John 17:15,Ephesians 6:13,James 4:7-8,Acts 14:22Second London Baptist Confession,Chapter 1, Paragraph 1, 10(V10) THE CONFIDENCE TO ENDURE SUFFERING1 Peter 1:5-7,2 Corinthians 4:16-18(V11) [AND] THE COMFORT OF HIS DOMINION.1 Peter 4:11b,Psalm 1:5-6,Hebrews 12:1b-2

Monday, April 20, 2026 - Week 17   CURE SYNGAP1 joins the Haystack Project in petitioning FDA for more clarity. PR: https://static1.squarespace.com/static/5966cc2220099e91326caaec/t/69d7dbb80155f46e144ae2e5/1775754168227/4.9.26+press+release+petition_vf.pdf Petition: https://static1.squarespace.com/static/5966cc2220099e91326caaec/t/69d7e016d5986e16ff09b64c/1775755288463/Letter+Head+Petition+for+rulemaking+to+amend%C2%A021+CFR+%C2%A7+314.126+and+%C2%A0%C2%A7312.47+%281%29.pdf Reuters 4/1: https://www.reuters.com/sustainability/boards-policy-regulation/rare-disease-advocacy-group-urges-trump-administration-restore-fda-clarity-2026-04-01/ Pink Sheet: https://insights.citeline.com/pink-sheet/pathways-and-standards/review-pathways/could-structured-not-ad-hoc-us-fda-flexibility-increase-rare-disease-development-certainty-BYXJENIJLFEINOSO72RS53ZJHE/   Show your support here: https://www.regulations.gov/document/FDA-2026-P-3666-0001 Paragraph 1 – Share information about you and SYNGAP1. Paragraph 2 – Seizures are hard to count while X, Y and Z are major burdens but FDA wants nice countable seizures, this makes it hard to develop drugs. Paragraph 3 – How could “clinically meaningful” endpoints potentially help your community and drug developers? How could a study design other than a ‘randomized clinical trial' help? How could FDA consulting with disease-specific experts help? Closing  – Finish your letter with anything along these lines: We support the framework for all rare set diseases in Haystack's petition. We don't believe FDA has to lower the evidentiary bar to approve treatments for our diseases. Randomization isn't always possible. New scientific methods should be considered. Endpoints specific to our disease should be considered. We urge FDA to open a rulemaking so we can have a legally binding regulation.   Board changes, thank you to everyone. https://www.linkedin.com/posts/curesyngap1_curesyngap1-syngap1-patientadvocacy-activity-7450528611339616256-4MJF?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAAD8f4B7JC4TMss45Q8hrsq5kiceI0Z8HE Press Release cureSYNGAP1.org/PR45   US, use your ICD-10, F78.A1: https://onlinelibrary.wiley.com/doi/10.1002/epi.70142   Study list! Citizen Health https://www.citizen.health/ai-advocate/syngap1 Combined Brain (May 1 & 2 in NorCal), https://docs.google.com/presentation/d/1IjaHILXj7AlBDlbTJgvYrkBS_0bnI8VCnTIiPXJ7JGM/edit?usp=sharing  ProMMiS https://www.linkedin.com/feed/update/urn:li:activity:7450196488300728320  Rare-X, the same week. DSC and Cook's are coming soon! CURE-ID for Drug responses.   CURE-ID is cool. https://cure.ncats.io/home (Webinar coming) Webinar: Thu May 7, 2026 1:30pm – 3pm (PDT) cureSYNGAP1.org/cureID    6th ANNUAL SPRINT FOR SYNGAP1, EVERYWHERE – 5 days - $207k! Go Tavilla. 17 teams raised $265K last year; this year, we have 20+ teams! https://curesyngap1.org/calendar/sprint4syngap-2026 Thank your Rifton for the donation of the tricycle. Email today: https://mailchi.mp/curesyngap1.org/sprint-for-syngap-2026-one-community-one-goal?e=17610baa03   INAUGURAL SF NIGHT OF IMPACT, CA – 38 days Join us this is our only Gala for 2026! cureSYNGAP1.org/SF26   5TH SCRAMBLE FOR SYNGAP, SC – 166 days  Classic case of a small event becoming an institution! cureSYNGAP1.org/Scramble26 PUBMED Pubmed 2026 is at 26. +9 vs the week. (61 last year was +9) https://pubmed.ncbi.nlm.nih.gov/?term=syngap1&filter=years.2026-2026&sort=date   SOCIAL MATTERS 4,891 LinkedIn.  https://www.linkedin.com/company/curesyngap1 1.55k YouTube.  https://www.youtube.com/@CureSYNGAP1 11.1k Twitter https://twitter.com/cureSYNGAP1 45k Insta https://www.instagram.com/curesyngap1   $CAMP closed at $4.67 Friday. https://www.google.com/finance/beta/quote/CAMP:NASDAQ   Like and subscribe to this podcast wherever you listen.  https://curesyngap1.org/podcasts/syngap10 Episode 205 of #Syngap10 #CureSYNGAP1 #Podcast

Easter is the preeminent feast in the Church's calendar. It celebrates the Resurrection of Jesus and foreshadows our new life in Christ and our opportunity for salvation. As such, it is appropriate to celebrate this feast for the entire 50 days long season the Church provides us. This blogcast explores “How to be an Easter People" from the Ad Infinitum blog, written by Erin Donn and read by Fatima Monterrubio Cruess.In St. Peter's speech at Pentecost he proclaims, “But God raised him up, releasing him from the throes of death, because it was impossible for him to be held by it.” (Acts 2:24) The truth of St. Peter's words is made manifest in the liturgical rhythm of the Church year. Just as our Lord could not be held by the grave, the celebration of Easter cannot be held to just one day. In her beauty, the Church does not stop at the Octave of Easter either, but offers the faithful an entire liturgical season—50 whole days—to glory in Jesus' resurrection. As Pope St. John Paul II declared, “We are an Easter people.” But once the beautiful liturgies of the Octave are behind us, the Easter lilies are removed from the churches, and the carrot cake has all been eaten, how are we to be an Easter people?One good place to begin is entering into the celebration of the liturgy. I have a distinct childhood memory of attending Easter Sunday Mass while my family was away on spring break. During the homily the priest invited everyone to exclaim ‘alleluia!', and then do it again because we weren't enthusiastic enough. As an unobservant and not-the-best catechised child, the purpose of this exercise was lost on me. But now, as a frequent daily Mass attendee I am much more aware of the forty long, alleluia-less days of Lent and more fully appreciate the significance of proclaiming alleluia together with renewed vigor on Easter and in the days that follow. During Lent many of us are motivated to make more time for Mass or prayer in our lives, which are beautiful practices to continue through the Easter season and beyond. Throughout the Masses of the Octave, the Victimae paschali laudes, an ancient chant, may be recited before the Gospel. Whether you attend Mass and hear the chant or simply take time to meditate on it each day, it is a beautiful tool for reflecting on the Paschal mystery in our lives and offering extra praise to the Lord: “The sheep are ransomed by the Lamb; / and Christ, the undefiled, / hath sinners to his Father reconciled.” Just as Lent is a time of remembering our mortality, Easter can be a time of remembering our Eternal life.Another liturgical tool to draw on is the Liturgy of the Hours, the communal prayer of the Church. In Morning Prayer throughout the Octave, we use the same psalms and canticle of Easter Sunday morning for all eight days. Particularly striking is the Canticle of Daniel (Daniel 3:57-88, 56), in which we pray that all things on Heaven and Earth bless the Lord:Let us bless the Father, and the Son, and the Holy Spirit.Let us praise and exalt him above all forever.Blessed are you, Lord, in the firmament of heaven.Praiseworthy and glorious and exalted above all forever.Whether you are able to pray Morning Prayer, or any of the other hours, in community or humbly pray them alone, you are still raising your prayers with all the faithful throughout the Earth and elevating them through Jesus to the Father.Throughout the Easter season, the first readings at Mass are taken from the Acts of the Apostles. Practically this makes sense since Acts recounts the events following Jesus' death and resurrection. But there is also a spiritual significance to reading from this book during the Easter season. Just like His first disciples, we aren't meant to keep our alleluias and joy in the Lord's resurrection and salvation confined to our churches or to ourselves, but to proclaim them to the world. It took a little encouragement for the apostles to do this, and so it may for us, but we too can say with St. Peter: “He commissioned us to preach to the people and testify that he is the one appointed by God as judge of the living and the dead. To him all the prophets bear witness, that everyone who believes in him will receive forgiveness of sins through his name.” (Acts 10:42-43) The Easter season is a grace-filled time to unite ourselves to this age-old mission of the Apostles and share the Good News of Jesus. The Catechism of the Catholic Church tells us that, “Beginning with the Easter Triduum as its source of light, the new age of the Resurrection fills the whole liturgical year with its brilliance.” (Paragraph 1168) At the Easter Vigil, the faithful are invited to renew their baptismal promises and remember the light of Christ that is within them. As we make our way through the Easter season, let us glorify Christ in the liturgy and let the brilliance of the Eternal Light shine forth from our churches and our hearts. Author:Erin Donn serves as the parish missionary at Immaculate Conception Church in Washington, DC. Resources:Listen to On Mission podcast Catholic Feast Days AppRead the Ad Infinitum blog Follow us:The Catholic Apostolate CenterThe Center's podcast websiteInstagramFacebookApple PodcastsSpotify Fr. Frank Donio, S.A.C. also appears on the podcast, On Mission, which is produced by the Catholic Apostolate Center and you can also listen to his weekly Sunday Gospel reflections. Follow the Center on Facebook, Instagram, X (Twitter), and YouTube to remain up-to-date on the latest Center resources.

Episode 165. Building of Relationship with Carolyn Kohlhaas  “If you, O Lord, should mark iniquities, Lord, who could stand? But there is forgiveness with you, so that you may be revered.”   Psalm 130: 3-4    Submit a Podcast Listener Question HERE!    Podcasts by Series  Level One Book Study  Level Two Book Study  "Moral existence is a response to the Lord's loving initiative. It is the acknowledgement and homage given to God and a worship of thanksgiving. It is cooperation with the plan God pursues in history." Catechism of the Catholic Church, Paragraph 2062    “To teach morality, the catechist must be convinced that at its roots is love, and then proceed from that principle. Morality emerges from love. We must never wander far from this source, not even when we teach explicitly the Law, which shows us the way we must follow to reach God. The entire Law has been summarized by Jesus in the two commandments of love, and He has made love the badge of allegiance. ‘One cannot obey more than one loves,' wrote the prior of the Little Brothers of Jesus.” (Teaching Doctrine and Liturgy)    “We can never stress too much the necessity of anchoring every moral teaching in a solid, kerygmatic orientation. Before providing specific instructions of what I am to do, moral formation must tell me with whom I do it, for whom I do it, with whom I am in relationship. Often we rush to action without having given enough thought to these primary questions.” (RPCII, 83)  “The primary and fundamental attitude of the person educated in the moral life is the awareness of belonging: to belong to God, to belong to God's world, to belong to humanity.”- Father Mongillo  Carolyn joins the conversation today to walk us through how morality evolves and progresses in a child's life and how we can meet those needs in all levels of the atria.  Carolyn Kohlhaas is a nationally recognized Formation Leader for CGS at all three levels of formation. She received her Bachelor of Arts Degree in Theology and in Religious Education from Franciscan University of Steubenville and her AMI Primary (3-6) and Elementary (6-12) Level Certifications from the Montessori Training Center of Minnesota. Carolyn has a Master of Education degree (Loyola College, Maryland) as well as a Master of Arts in Theology (St. Paul Seminary School of Divinity). She has worked with children in Atriums in the Archdiocese of St. Paul and Minneapolis since 2006 and with adults in formation courses across the US since 2009.  wosatrium.weebly.com      Books you might be interested in:  Development of the Rome Atrium  The Religious Potential of the Child 6 to 12 Year Olds  History of the Kingdom of God Part 1: Creation to Parouisa  History of the Kingdom of God Part 2: Liturgy and the Building of the Kingdom  Life in the Vine: The Joyful Journey Continues    Podcast Episodes you might be interested in:  Episode 34 – The Paschal Mystery  Episode 55 – Cosmic Education      TINA LILIG MEMORIAL FUND  During the month of February, we remember our first national director, Valentina Lillig, and we honor her legacy by encouraging generosity to the Tina Lillig Memorial Fund.     The mission of the Tina Lillig Memorial Fund is to support the growth and development of Catechesis of the Good Shepherd through catechist formation, missionary outreach, and the work of the United States Association of The Catechesis of the Good Shepherd (CGSUSA). To these ends, the TLM Fund offers a source of financial aid (in the form of partial scholarships) to applicants who would find paying the full tuition of a course to be a hardship.    If you or your community needs scholarship or grant support for formation, please visit our website and apply [this could be a link in the show notes]. The next deadline for applications is April 30. Thank you to all the donors who have generously supported the TLM Fund and made formation courses possible for catechists across the US and the world.      BECOME A CGSUSA MEMBER          AUDIOBOOK:    Audiobook – Now Available on Audible  CGSUSA is excited to offer you the audio version of The Religious Potential of the Child – 3rd Edition by Sofia Cavalletti, read by Rebekah Rojcewicz!  The Religious Potential of the Child is not a “how-to” book, complete with lesson plans and material ideas. Instead it offers a glimpse into the religious life of the atrium, a specially prepared place for children to live out their silent request: “Help me come closer to God by myself.” Here we can see the child's spiritual capabilities and perhaps even find in our own souls the child long burdened with religious information. This book serves as a companion to the second volume, The Religious Potential of the Child 6 to 12 Years Old. The desire to have this essential text available in audio has been a long-held goal for many. The work of many hands has combined to bring this release to life as an audiobook.    Find out more about CGS:       Learn more about the Catechesis of the Good Shepherd     Follow us on Social Media-  Facebook at “The United States Association of the Catechesis of the Good Shepherd”  Instagram-  cgsusa  Twitter- @cgsusa  Pinterest- Natl Assoc of Catechesis of the Good Shepherd USA  YouTube- catechesisofthegoodshepherd 

Pastor Mark Redfern | Series: The Obedience of Faith

In this week's episode, we take a look at eight common ebook formatting errors and how to correct them. This coupon code will get you 25% off the ebook of Ghost Exile: Omnibus One at my Payhip store: EXILE25 The coupon code is valid through April 6, 2026. So if you need a new ebook this winter, we've got you covered! TRANSCRIPT 00:00:00 Introduction and Writing Updates Hello, everyone. Welcome to Episode 296 of The Pulp Writer Show. My name is Jonathan Moeller. Today is March 27th, 2026 and today we are looking at eight ebook formatting problems that readers hate and how to fix them. Before we get into our main topic, we'll have Coupon of the Week and a progress update on my current writing, publishing, and audiobook projects. So let's start off with Coupon of the Week. This week's coupon code will get you 25% off the ebook of Ghost Exile: Omnibus One at my Payhip store. That coupon code is EXILE25. And as always, you can get the coupon code and the links to my Payhip store in the show notes. This coupon code will be valid through April 6th, 2026. So if you need a new ebook as we leave winter and head into spring, we have got you covered. Now for an update on my current writing, publishing, and audiobook projects. As of this recording, I am currently 97,000 words into Blade of Wraiths, and I'm hoping to get to the 100,000 word mark by the end of today. You will note that if you read the first three books in the series, this will make it longer than the first three by good bit, which is part of the reason why it's taking so long. I was hoping to have it published by now, to be honest, but between the length and the variety of things I've had to do in real life before I can get to writing have slowed me down a bit. I have made progress and I am hoping to finish the rough draft before the 31st, if all goes well. Then it will be time to write a tie-in short story that newsletter subscribers will get for free when Blade of Wraiths is published, and then on to editing. So I'm hoping to have the book out in the second half of April, if all goes well and the creek doesn't rise, so to speak. I'm also 12,000 words into Dragon Mage, which will be the sixth book in the Rivah Half-Elven Thief series. I'm hoping that will be out in May because that will become the main project once Blade of Wraiths is finished. In audiobook news, as I mentioned before, recording of Blade of Storms (as excellently narrated by Brad Wills) is complete. I think as of right now, you can get it at my Payhip store, Google Play, and Kobo. Audible, Amazon, Apple, and a few of the other major stores should hopefully be coming along before too much longer. Recording is done on Wizard Assassin, which you may recall is the fifth book in the Rivah series, and that is narrated by Leanne Woodward. We just have to proof that, and then it'll be time to submit it for processing at the stores. So hopefully you'll be able to get your hands on that before too much longer. I believe next week, Hollis McCarthy will start recording on Cloak of Illusion, the 12th book in the Cloak Mage series. Since a few people have asked about this, I thought I'd mentioned here there will be a Cloak Mage: Omnibus Four in audio that will combine Cloak of Embers, Cloak of Titans, and Cloak of Illusion. I will probably start working on that one month after Cloak of Illusion comes out in audio. So that will probably be along sometime this summer, if all goes well. So that's where I'm at on my current writing, publishing, and audiobook projects. We've got good things coming up for you soon. 00:03:18 Main Topic: ebook Formatting That Readers Hate Now let's move on to our main topic this week, and it's something I have quite a bit of experience with, I have to say, is ebook formatting that readers hate. Today, we're going to talk about issues with ebook formatting, a topic that people have very strong opinions about. I am going to talk about eight issues in particular and then discuss how we can fix or prevent or best of all, avoid them. First of all, what do we mean by ebook formatting? Formatting is the term to describe the layout, text, and images of the book. Back in the print book days, this was a complicated but fairly fixed thing. In the very old days, you would have to lay it out the books manually. A few decades ago, they had programs like QuarkXPress and Adobe InDesign where you could use these software applications to prepare finished files for sending to the printer to be printed as books. But nowadays, in an age where books are electronic files read on dozens of possible devices and apps, it's much more difficult to predict every possible thing that can go wrong with formatting. For example, with a print book in the old days, you would set the layout, and that is the way the layout would look. Every single book (ideally) printed would look the same. But in the modern age, if you have an ebook, it could be read on a Kindle, a Kindle Color, the Kindle app on the phone, a Barnes and Noble Nook, a Google tablet, on Apple iBooks on an iPod, on Apple iBooks on an iPhone, on the Kindle app on an iPhone. There are literally dozens of different potential combinations where an ebook file could be read and therefore dozens of different potential complications that can arise for ebook formatting. If there are formatting problems, readers will not be able to understand or even physically be able to read the text at worst, and more likely will be deeply annoyed by the issues that feel like road bumps such as extra line breaks and will express their displeasure in reviews. There are also an abundance of ebooks out there that are hastily converted from PDF or Word docs that make for an extremely difficult reading experience. Many of the problems we're going to talk about today arise from when that happens, including the first several. So with that in mind, here are eight ebook formatting errors to avoid and how to deal with them. #1: Illustrations, charts, and maps. Some of the most common issues are that these aren't legible in grayscale (like an e-ink reader, like the Kindle or the various Kobo devices) or have too small fonts or a poor quality or low resolution. This is a very common problem, and it's kind of a problem across all ebooks. Like for example, think of the map of Middle Earth from J.R.R. Tolkien's The Lord of the Rings. It's a very complicated map with a lot of different names on it. Now imagine trying to read that map on a six inch black and white Kindle screen. You probably aren't going to be able to do it very well. People using dark mode are most likely going to have problems with seeing images in the same way. So if you have an image heavy book, it's a good idea to test that. So how to fix this? What I've done myself, because I write a lot of fantasy novels and fantasy novels traditionally need maps, is I will put the full color maps on my website and then in the author's note at the beginning, say a map of Owyllain or a map of Andomhaim is available at the author's website on this link. I had a few people ask if I could put the maps in the books and the nearest thing I've done is including the maps with some of the direct sales of my Payhip store. I think the system works very well because those map pages are consistently some of the most visited pages on my website. And to be honest, it would be a lot easier to look at the map on, for example, your tablet or your computer instead of on your phone when you're trying to use the e-reader application. And even if you are looking at the map on my website on your phone, you can pinch zoom in the browser in the way you probably can't in the ebook reader. So that is the quick and easy solution for it. For more image intensive books like textbooks, for example, you just have to take great care and make sure they use large, high resolution images that aren't going to get pixelated. It may be a good idea to set up a website that has these images that the readers can refer to. This would be a bit of extra expense and work, obviously, but it would go a long way to making your readers happy, like I have done with the maps on my website. #2: Paragraphs that aren't indented or unneeded spaces between indented paragraphs. This is one of the biggest pet peeves that people have when reading ebooks. Indented paragraphs are important and if they're properly indented, then space isn't needed between the paragraphs and this will look odd to readers. This was in fact a problem I did have the first couple of years of my ebook publishing career because I used Sigil for my primary ebook formatting and that is the default way Sigil renders ebooks. In time, I started to switch to using Vellum for ebook formatting and that solved the problem. So what I did at that point was I just went through all my entire library, reformatted the ebooks in Vellum, and then re-uploaded them to the various ebook store platforms and that took care of the problem. And that's what I've been doing ever since. I've been using Vellum for ebook formatting since at least 2018 at this point, I think. So how to fix? The easiest way to fix this is to spend a little money on a tool like a Vellum or Atticus or a similar program that does ebook formatting for you because that will make sure your paragraphs are indented properly and then you can spot check as well to make sure that the table of contents is working and the chapters are working and that the paragraphs are being indented properly. #3: Our third issue is hard to read fonts. Some people enjoy fonts that are very stylized, but let's be honest, most readers don't. It's also an accessibility issue if the serifs or the flourishes are too elaborate. How to fix? Honestly, the easiest way is to not use custom fonts for your book unless you have a really, really, really and I emphasize really good reason to do so because part of the appeal of ebooks and ereaders is that you can adjust the font size to whatever you want, which is also particularly important as people get older and it becomes harder to read small print. So if you have a font that is inherently difficult to read in your book, that is going to be a serious, serious speed bump to readers enjoying the book. So my advice for custom fonts is avoid if possible, and even if not possible, still try to avoid. #4: Our fourth issue is basically the same as the third issue, except worse. Font sizes that can't be changed/hard coded text size. A bad font choice is made much worse if the ebook is hard coded to a certain font. This can happen if you use certain exporting tools to export your print book layout into an ebook layout, and sometimes the font size carries over with that. People have strong feelings about fonts and some are more accessible than others, so it's very important not to do this. So how to fix this problem? Don't do it to begin with. It's bad and people will be mad at you. Do not hard set the font size in your ebook, and if you do it by accident, fix it if at all possible. That is my advice in this situation. #5: A fifth potential ebook formatting issue is random hyphens in the middle of words that shouldn't be there. Random hyphens in the middle of words can happen if you convert a book from a PDF file to an EPUB file, and this means it's a very preventable formatting issue. The best way to avoid this is not to upload a PDF file or a Word file to Kindle Direct Publishing and the other ebook platforms, because while they do have automated tools for converting a PDF and a Word document to an EPUB, there's often weird formatting glitches that pop up like the random hyphenations. The best way to avoid this is to use a proper tool to format your ebook like Vellum, Atticus, maybe Scrivener, maybe Kindle Create. So that is the best way to do that by preparing your own EPUB file and uploading that directly. #6: A sixth issue is footnotes or bibliography or endnotes that don't link back to the original page. This also really annoys people. Footnotes and other links within books either need to be popups or linked back to the original page otherwise, readers will be deeply annoyed trying to return to their place each time and even fewer people will read them. And the best way to fix this is to make sure you have set up the footnotes, bibliography, or endnotes properly in a formatting program so that the links work. If you try to do this using a PDF or Word document you upload to the ebook platforms, it's probably not going to work. So it's best to, again, to create your own EPUB and upload it directly. #7: A seventh thing that annoys readers is books without chapters or books without numbered chapters. Books without chapters or unnumbered chapters is a very annoying thing for readers, but most formatting software will help to create and number chapters. Now, how to fix? The technical explanation of this is an EPUB file is essentially a zipped version of several XHTML files and each XHTML file in the sort of zipped package should be its own chapter. The way an EPUB generates table of content files is that every chapter heading should have a H1 HTML key, and that it lets the EPUB know which heading should be chapters. Now you can hand code all this in a program like Sigil, but again, it's much easier to use a program like a Vellum, Atticus, or Kindle Create that will automatically create a proper table of contents for you rather than you having to go through with a XHML or EPUB editor like Sigil and then hand code the H1 tag. It's much easier to automate this and there are applications that will do this for you. #8: The eighth formatting problem that annoy readers is quotation marks the wrong way. What do we mean by that? Quotation marks come in two styles, straight and curly. Curly ones look like flipped commas, while straight quotation marks look more like two lines. Most word processing programs like Word will default the curly quotation marks and convert straight ones. Although curly quotations can sometimes create garbled formatting, they're considered to be the standard. How to fix? Honestly, the easiest way to fix this is to, if it's a problem for you, is to do a find and replace or a find and search. Just if you find that usually what happens is that if you hold down the shift key for too long while you're thinking and then you hit the quotation marks. Sometimes in certain operating systems, it'll convert it to straight quotation marks instead of curly ones. So if you are worried about a few of the wrong type of quotation marks showing up in your document, the easiest way to fix it is just a quick search through the document for straight quotation marks and then remove any ones that you see. We've talked a lot about formatting ebooks and tools for formatting ebooks. So I thought I would suggest the three most popular applications for formatting ebooks, and those would be Vellum, Atticus, and Kindle Create. Vellum is most expensive, but I think it has most features, whereas Kindle Create is pretty simple, but it's free. Vellum is Mac only. As I've mentioned before on the show, I started using it for print layout and then liked it enough that I switched over to using it for all my ebook formatting and that's been true for like the last eight years now. It's very easy to learn. It has excellent features and excellent support. I've never had any technical problems with it whatsoever. I'd say the biggest liabilities for it are that it's the most expensive of the options and it's Mac only. So if you don't have access to a Mac, you would need to choose another option. The next option would be Atticus, which the advantage for Atticus is that it's cross platform. It will work on Windows, Mac, and Linux because it is foundationally a web application. It's also very collaborative in that you can invite other users to use it. The creator of the program, the guy who runs Kindlepreneur named Dave Chesson, has said he wants to create an all- in-one ebook creation, writing, and formatting tool with Atticus, and I think he's taken a pretty good stab at it. I'd say the biggest weakness for Atticus is that it does have a bit of a learning curve because there are so many different functions packed into it. The free option would be Kindle Create, which is a program put out by Amazon. Kindle Create is pretty basic, but it does have everything you need and you can use it to create both files optimized for the Kindle platform and generic EPUBs that you can use on other platforms. I'd say its biggest weakness is that it's fairly simple and that it's produced by Amazon, which some people have strong feelings about, but it is free, which is hard to beat. In Episode 251 (about a year ago), I talked about ebook formatting software and offered pros and cons for each. There isn't one perfect option, so it's good to look over the details for each one of them and watch one of the many available reviews or overviews available online to get a feel for how each one works differently. And now the three strategies for finding formatting issues before your readers do. #1: Test your ebook in each file format (PDF/EPUB) and on multiple devices. It's impossible and prohibitively expensive to test on every possible device, but try some of the most common devices and apps. It's important to test on a Kindle, an actual Kindle and not just on the app, for example. #2: Check the page and chapter breaks in particular. These are the spots where odd formatting tends to cluster. Check all of them to make sure there's nothing weird that you need to fix. #3: Check for legacy things from the print version like page numbers, the table of contents, and so forth. And I find that the easiest way to do all three of these at once is in my final phase of editing a book, I will have my computer read it aloud to me with text to speech and I find that's a good way to find any remaining typos. It's also a very good way of spotting formatting issues since you'll be looking over the entire book one more time and if there's like a weird gap or a line break or something of that nature, you're probably going to have a good chance of spotting it. So in conclusion, ebooks can be a bit difficult to format, but there is dedicated software and a slew of tutorials and guides to help you do it. Learning a little bit of HTML and CSS can help, but it's not a requirement at this point. If you're self-publishing, having a well formatted book is important because readers either won't enjoy or won't even start a poorly formatted ebook. I hope these tips help you with knowing where to start with improving your ebook formatting. So that is it for this week. Thank you for listening to the Pulp Writer Show. I hope you found the show useful. A reminder that you can listen to all the back episodes on https://thepulpwritershow.com. If you enjoyed the podcast, please leave your view on your podcasting platform of choice. Stay safe and stay healthy and we'll see you all next week.  

Send us Fan MailHeather Binning is joined by Michael Foran, Associate Professor of Law at the University of Oxford and Tutorial Fellow at Keble College, specialising in public law, equality, and anti-discrimination law.Drawing on cases including Sandie Peggie and the Good Law Project's latest challenge to EHRC guidance, Michael cuts through the legal complexity to explain — in plain language — what these developments mean and how they are being weighed by legal professionals.He also offers practical advice for anyone facing discrimination from an employer, or being refused service by event organisers as a gender-critical individual or group.Michael's first book, Equality Before the Law (Hart, 2023), won the Yorke Prize from Cambridge. His second book, Sex, Gender Identity, and the Law, is forthcoming from Cambridge University Press in 2026.===Useful Links:Michael's Substack on the GLP appeal:  https://knowingius.orgEquality Act Schedule 3, Paragraph 26: https://www.legislation.gov.uk/ukpga/2010/15/schedule/3/paragraph/26/enacted/data.xht?view=snippet&wrap=trueThe Public Sector Equality Duty (PSED): https://www.gov.uk/government/publications/public-sector-equality-duty-guidance-for-public-authorities/public-sector-equality-duty-guidance-for-public-authoritiesIf you enjoyed listening to this podcast and would like to hear more, go to womensrights.network/wrn-podcast to listen, download and subscribe to more of our episodes.  And if you'd like to join our conversations, go to womensrights.network/join-wrn

Second London Baptist Confession of Faith, Chapter 3, Paragraph 7

15: Bigwig's Character and the Lost Paragraph 15.1 Burrowkeeping I made my favourite mistake yet last week, when I referred to 'The Warren of the Snares' as 'The Warren of the Swears' That would make for a completely different story. One where they have to leave Cowslip's Warren because Fiver has an issue with all the bad language. Which is a bit rich for a rabbit who comes out with such potty-mouthed stuff as "embleer Frith". 0:47​15.2: Bigwig's Character and the Lost Paragraph revisited. 7:35​15.3 The Lost Paragraph analysed 13:03​15.4 Next episode Next week I return to going through the book and we begin Part Two with perhaps the pivotal Chapter of the entire story. A chapter with the title...Watership Down.

8: Chapters 10 and 11: The Road and the Common and Hard Going 8.1 Housekeeping: I referred briefly to Speedwell waking up Acorn It was Hawkbit waking up Speedwell only. Bigwig's phrase "u embleer rah" suggests a more complex meaning of "u", possibly meaning "you" as well as "the" depending on context. Interesting discussion of the book in the latest episode of the 'Read it twice' podcast on spotify. Episode 2. There is now a link in the description if you would like to leave a voice message for possible use in the podcast. The link is https://anchor.fm/wate...​ 3:36​8.2 Chapter 10. The Road and the Common. 9:08​8.3 Chapter 11. Hard Going and The Lost Paragraph. 18:22​8.4 Next Episode The next episode covers the beginning of the last 6 chapters of the end of Part One of the book. These chapters cover that metaphorical mist that Fiver spoke of in trance when he first saw the distant Watership Down. Vocab: Embleer: Stinking (missed in last episode) Hrududu: a car or other human vehicle Yona: Hedgehog

Episode Notes Andrew sits down with writer and disability activist Sloan Harlow for an honest, unfiltered conversation about navigating polyamory, kink, dating, grief, and joy as a disabled person. Sloan opens up about living with hypermobile Ehlers-Danlos syndrome, bipolar one disorder, and a constellation of other conditions, sharing wisdom on solo polyamory, accessible kink, medical ableism, and why the disability community needs more grit + so much more! Follow them here: www.instagram.com/disabledparagraphs Episode Sponsors Do you wanna turn b*tt stuff up a notch. Go to bvibe.com and use code AFTERDARK to receive 20% off orders of $100 (including bundles, discounted items and more). Disability content creation doesn't have to be hard. Follow @seated.perspectives on Instagram to learn how to make content creation a gentle, easy, accessible experience. Are you looking for attendant care when you need it at your convenience? Check out your team, on tap www.whimble.ca Get 15% off your next purchase of sex toys, books and DVDs by using Coupon code AFTERDARK at checkout when you shop at trans owned and operated sex shop Come As You Are www.comeasyouare.com Order Notes From a Queer Cripple and hire him to speak on it by e-mailing andrew@andrewgurza.com US: https://us.jkp.com/products/notes-from-a-queer-cripple Canada: https://www.ubcpress.ca/notes-from-a-queer-cripple Support the show with a donation: https://patreon.com/disabilityafterdark This podcast is powered by Pinecast.

Pastor David Preaches from 1 Peter 3:18-22.Outline(v. 18A) Christ's Victory Accomplished1 Peter 3:17Philippians 1:291 Peter 4:1aHebrews 9:26bHebrews 10:14Hebrews 7:27Romans 5:1-2John 19:30(vv. 18b-20) His Victory ProclaimedActs 17:11Colossians 2:15Acts 2:23-24, 25-27 (Psalm 16:8-11Genesis 6:2, 4; 7:21-232 Peter 2:4-5, 9Jude 1:6The Apostle's Creed, Paragraph 2(vv. 21-22) and His Victory AppliedColossians 2:17Romans 6:3-4Galatians 3:27Galatians 2:20Romans 6:3-4The Baptist Catechism, Q&A #96

Second London Baptist Confession of Faith, Chapter 3, Paragraph 6

NUMBER FILE SEPARATE SINGLE PARAGRAPH 1-12 1. Mary Anastasia O'Grady of the *Wall Street Journal* explores the presence of Iranian assets in Venezuela, noting that Iranian technology was used to build an assembly plant for military drones in the state of Aragua. She discusses the status of acting president Delcy Rodriguez, whom the U.S. recently recognized as the sole head of state in federal court, a move that appears to sideline the democratic opposition leader Maria Corina Machado. O'Grady also highlights the legal significance of Alex Saab, an interlocutor between Tehran and Caracas, who is viewed as a "treasure trove" of information regarding the criminal charges against the Maduro regime. (1)1900

Today's poem is Five Paragraph Essay on Time by Kathleen Flenniken. The Slowdown is your daily poetry ritual. In this episode, Maggie writes… “Recently, I told a friend that I had procrastinated a task, and so I had to really hustle to get it done on time, and she kindly corrected me. Or, rather, she reframed what I was calling procrastination as something else: triage. That's what she called it. She said, “You have so much to do, you have to triage tasks—tackle the big and immediate ones first, and let some of the smaller ones go for a bit.” She had a point. I didn't have a time management issue or a lack of focus. I was juggling multiple tasks, and that meant that some of them naturally had to wait.” Celebrate the power of poems with a gift to The Slowdown today. Every donation makes a difference: https://tinyurl.com/rjm4synp

Pastor David preaches from 1 Peter 3:8-12.Outline(vv. 8-9) The conduct of suffering saints;Romans 12:14-182 Corinthians 13:112 Timothy 3:16-17Romans 15:5-7Romans 12:15Galatians 6:2Hebrews 13:1John 13:34-35Ephesians 4:32Matthew 9:36Mark 6:34Philippians 2:3Luke 6:281 Corinthians 4:12bSecond London Baptist Confession: Chapter 1, Paragraph 1(V. 9b) Called to obtain a great blessing1 Peter 1:3-51 Peter 3:9Matthew 5:10-12(vv. 10-12) and God's favor towards godly sufferersPsalm 34:12-161 Samuel  21-22Ecclesiastes 2:17, 25Psalm 141:3-4a2 Corinthians 12:9-10Psalm 37:7-9

Send a textReady for a pivot that changes how you relate to God, Jesus, and your own mind? We unpack why A Course in Miracles starts by training perception before speaking directly about God—and how that protects us from turning the divine into another ego defense. We explore the counterintuitive teaching that awe belongs to revelation alone, not to teachers, and why equality with Jesus is the necessary ground for learning rather than a flattering slogan.Together we clarify Jesus' precise role in the atonement: coordinating, not replacing your choice; performing miracles indiscriminately because there's no personal investment to distort them; and holding the template of a healed mind within perception. We contrast this with our role—to accept the atonement for ourselves and extend it—so miracles flow as natural, involuntary expressions of remembered innocence. You'll hear how ACIM reframes projection into true extension, how refusing hierarchy keeps power inside where it belongs, and how guidance becomes simple when we stop making differences real.We also draw a sharp line between therapeutic reframing and genuine undoing. Psychology can reorganize the story; ACIM dissolves the premise. That shift shows up in daily choices: listening over strategizing, inclusion over grievance, efficiency over struggle. Along the way we place the atonement outside of time, note the many minds that have demonstrated it, and keep God beyond perception while welcoming help within it.If you've ever felt torn between worship and skepticism, this conversation threads the needle: gratitude without idolatry, rigor without rigidity, and a practice you can use now—accept guiltlessness for everyone, often and without exception. Subscribe, share with a friend, and leave a review with your biggest takeaway so we can keep exploring the Course with clarity and heart.Support the show

What if a fifth grader could turn a pile of ideas into a clear, grounded paragraph—every time—without leaning on AI? We bring back dyslexia researcher Russell Van Brocklen for part three of our series to show exactly how: start with a hero, a universal theme, and a villain; distill three good reasons into one-word themes; and anchor everything to a real quote. The result is a body paragraph that's honest, teachable, and repeatable—plus a writing process students can explain step by step.We also address integrity in the AI era: students must show their process or redo the work, then later use AI as a research coach rather than a shortcut. By the end, you'll have a clear path to scale from one body paragraph to three, then add a thesis and conclusion that help students pass state tests and feel proud of their writing.If this helped, follow the show, share it with a colleague, and leave a review. Grab the free resources in the notes, and stay tuned for part four, where we break down concrete, classroom-ready examples.

Send a textJoin Drake and Marie as they explore how to write interiority using a 12 layer model. We go over everything from showing, to telling, to worldview and more! Writer's room (50% off for lifetime membership): https://writersroom.mn.co/plans/338439?bundle_token=196fd3965307a65eee0d1bf2bc6fa5a6&utm_source=manualMembership for Just In Time Worlds: https://www.youtube.com/channel/UCxvBH0EkwuHsQ9ryHHQNi2Q/joinGive us feedback at releasingyourinnerdragon(at)gmail(dot)comDiscord: https://discord.gg/vMrmBsF5fhMagicfall: http://magicfallnovel.com/ReleasingYourInnerDragon@gmail.comDrake's Contact Details:Starving Writer Studio: https://www.starvingwriterstudio.com/Drake-U: https://class.drakeu.com/  - Use RYID25 for 25% off!Writer's Room: https://writersroom.mn.co/Scavenger hunt: https://www.starvingwriterstudio.com/scavenger-huntMarie's contact details:Books: https://mariemullany.com/booksJust In Time Worlds: https://www.youtube.com/c/JustInTimeWorlds?sub_confirmation=1

Leon Morris once said that Romans 3:21-27 is the most important paragraph ever written. The reason that he said this is captured well by what Charles Spurgeon said of this passage. He said "I do not know any passage in Scripture in which the way of salvation is more clearly and fully set forth than this one." John Piper said about this passage that, "if you understand this passage, you understand Christianity." Martin Luther said that this passage is "the central place of the book of Romans, and of the entire Bible." Nowhere is the gospel so clear. Nowhere is the logic of salvation so powerfully set forth as it is in this passage. In this passage we see that salvation is by Grace alone, through Faith alone, in Christ alone, and to the glory of God alone. This episode is a sermon that was originally preached by Jackson Hankey at a Christian Retreat for College Athletes.

From rewriting Google's search stack in the early 2000s to reviving sparse trillion-parameter models and co-designing TPUs with frontier ML research, Jeff Dean has quietly shaped nearly every layer of the modern AI stack. As Chief AI Scientist at Google and a driving force behind Gemini, Jeff has lived through multiple scaling revolutions from CPUs and sharded indices to multimodal models that reason across text, video, and code.Jeff joins us to unpack what it really means to “own the Pareto frontier,” why distillation is the engine behind every Flash model breakthrough, how energy (in picojoules) not FLOPs is becoming the true bottleneck, what it was like leading the charge to unify all of Google's AI teams, and why the next leap won't come from bigger context windows alone, but from systems that give the illusion of attending to trillions of tokens.We discuss:* Jeff's early neural net thesis in 1990: parallel training before it was cool, why he believed scaling would win decades early, and the “bigger model, more data, better results” mantra that held for 15 years* The evolution of Google Search: sharding, moving the entire index into memory in 2001, softening query semantics pre-LLMs, and why retrieval pipelines already resemble modern LLM systems* Pareto frontier strategy: why you need both frontier “Pro” models and low-latency “Flash” models, and how distillation lets smaller models surpass prior generations* Distillation deep dive: ensembles → compression → logits as soft supervision, and why you need the biggest model to make the smallest one good* Latency as a first-class objective: why 10–50x lower latency changes UX entirely, and how future reasoning workloads will demand 10,000 tokens/sec* Energy-based thinking: picojoules per bit, why moving data costs 1000x more than a multiply, batching through the lens of energy, and speculative decoding as amortization* TPU co-design: predicting ML workloads 2–6 years out, speculative hardware features, precision reduction, sparsity, and the constant feedback loop between model architecture and silicon* Sparse models and “outrageously large” networks: trillions of parameters with 1–5% activation, and why sparsity was always the right abstraction* Unified vs. specialized models: abandoning symbolic systems, why general multimodal models tend to dominate vertical silos, and when vertical fine-tuning still makes sense* Long context and the illusion of scale: beyond needle-in-a-haystack benchmarks toward systems that narrow trillions of tokens to 117 relevant documents* Personalized AI: attending to your emails, photos, and documents (with permission), and why retrieval + reasoning will unlock deeply personal assistants* Coding agents: 50 AI interns, crisp specifications as a new core skill, and how ultra-low latency will reshape human–agent collaboration* Why ideas still matter: transformers, sparsity, RL, hardware, systems — scaling wasn't blind; the pieces had to multiply togetherShow Notes:* Gemma 3 Paper* Gemma 3* Gemini 2.5 Report* Jeff Dean's “Software Engineering Advice fromBuilding Large-Scale Distributed Systems” Presentation (with Back of the Envelope Calculations)* Latency Numbers Every Programmer Should Know by Jeff Dean* The Jeff Dean Facts* Jeff Dean Google Bio* Jeff Dean on “Important AI Trends” @Stanford AI Club* Jeff Dean & Noam Shazeer — 25 years at Google (Dwarkesh)—Jeff Dean* LinkedIn: https://www.linkedin.com/in/jeff-dean-8b212555* X: https://x.com/jeffdeanGoogle* https://google.com* https://deepmind.googleFull Video EpisodeTimestamps00:00:04 — Introduction: Alessio & Swyx welcome Jeff Dean, chief AI scientist at Google, to the Latent Space podcast00:00:30 — Owning the Pareto Frontier & balancing frontier vs low-latency models00:01:31 — Frontier models vs Flash models + role of distillation00:03:52 — History of distillation and its original motivation00:05:09 — Distillation's role in modern model scaling00:07:02 — Model hierarchy (Flash, Pro, Ultra) and distillation sources00:07:46 — Flash model economics & wide deployment00:08:10 — Latency importance for complex tasks00:09:19 — Saturation of some tasks and future frontier tasks00:11:26 — On benchmarks, public vs internal00:12:53 — Example long-context benchmarks & limitations00:15:01 — Long-context goals: attending to trillions of tokens00:16:26 — Realistic use cases beyond pure language00:18:04 — Multimodal reasoning and non-text modalities00:19:05 — Importance of vision & motion modalities00:20:11 — Video understanding example (extracting structured info)00:20:47 — Search ranking analogy for LLM retrieval00:23:08 — LLM representations vs keyword search00:24:06 — Early Google search evolution & in-memory index00:26:47 — Design principles for scalable systems00:28:55 — Real-time index updates & recrawl strategies00:30:06 — Classic “Latency numbers every programmer should know”00:32:09 — Cost of memory vs compute and energy emphasis00:34:33 — TPUs & hardware trade-offs for serving models00:35:57 — TPU design decisions & co-design with ML00:38:06 — Adapting model architecture to hardware00:39:50 — Alternatives: energy-based models, speculative decoding00:42:21 — Open research directions: complex workflows, RL00:44:56 — Non-verifiable RL domains & model evaluation00:46:13 — Transition away from symbolic systems toward unified LLMs00:47:59 — Unified models vs specialized ones00:50:38 — Knowledge vs reasoning & retrieval + reasoning00:52:24 — Vertical model specialization & modules00:55:21 — Token count considerations for vertical domains00:56:09 — Low resource languages & contextual learning00:59:22 — Origins: Dean's early neural network work01:10:07 — AI for coding & human–model interaction styles01:15:52 — Importance of crisp specification for coding agents01:19:23 — Prediction: personalized models & state retrieval01:22:36 — Token-per-second targets (10k+) and reasoning throughput01:23:20 — Episode conclusion and thanksTranscriptAlessio Fanelli [00:00:04]: Hey everyone, welcome to the Latent Space podcast. This is Alessio, founder of Kernel Labs, and I'm joined by Swyx, editor of Latent Space. Shawn Wang [00:00:11]: Hello, hello. We're here in the studio with Jeff Dean, chief AI scientist at Google. Welcome. Thanks for having me. It's a bit surreal to have you in the studio. I've watched so many of your talks, and obviously your career has been super legendary. So, I mean, congrats. I think the first thing must be said, congrats on owning the Pareto Frontier.Jeff Dean [00:00:30]: Thank you, thank you. Pareto Frontiers are good. It's good to be out there.Shawn Wang [00:00:34]: Yeah, I mean, I think it's a combination of both. You have to own the Pareto Frontier. You have to have like frontier capability, but also efficiency, and then offer that range of models that people like to use. And, you know, some part of this was started because of your hardware work. Some part of that is your model work, and I'm sure there's lots of secret sauce that you guys have worked on cumulatively. But, like, it's really impressive to see it all come together in, like, this slittily advanced.Jeff Dean [00:01:04]: Yeah, yeah. I mean, I think, as you say, it's not just one thing. It's like a whole bunch of things up and down the stack. And, you know, all of those really combine to help make UNOS able to make highly capable large models, as well as, you know, software techniques to get those large model capabilities into much smaller, lighter weight models that are, you know, much more cost effective and lower latency, but still, you know, quite capable for their size. Yeah.Alessio Fanelli [00:01:31]: How much pressure do you have on, like, having the lower bound of the Pareto Frontier, too? I think, like, the new labs are always trying to push the top performance frontier because they need to raise more money and all of that. And you guys have billions of users. And I think initially when you worked on the CPU, you were thinking about, you know, if everybody that used Google, we use the voice model for, like, three minutes a day, they were like, you need to double your CPU number. Like, what's that discussion today at Google? Like, how do you prioritize frontier versus, like, we have to do this? How do we actually need to deploy it if we build it?Jeff Dean [00:02:03]: Yeah, I mean, I think we always want to have models that are at the frontier or pushing the frontier because I think that's where you see what capabilities now exist that didn't exist at the sort of slightly less capable last year's version or last six months ago version. At the same time, you know, we know those are going to be really useful for a bunch of use cases, but they're going to be a bit slower and a bit more expensive than people might like for a bunch of other broader models. So I think what we want to do is always have kind of a highly capable sort of affordable model that enables a whole bunch of, you know, lower latency use cases. People can use them for agentic coding much more readily and then have the high-end, you know, frontier model that is really useful for, you know, deep reasoning, you know, solving really complicated math problems, those kinds of things. And it's not that. One or the other is useful. They're both useful. So I think we'd like to do both. And also, you know, through distillation, which is a key technique for making the smaller models more capable, you know, you have to have the frontier model in order to then distill it into your smaller model. So it's not like an either or choice. You sort of need that in order to actually get a highly capable, more modest size model. Yeah.Alessio Fanelli [00:03:24]: I mean, you and Jeffrey came up with the solution in 2014.Jeff Dean [00:03:28]: Don't forget, L'Oreal Vinyls as well. Yeah, yeah.Alessio Fanelli [00:03:30]: A long time ago. But like, I'm curious how you think about the cycle of these ideas, even like, you know, sparse models and, you know, how do you reevaluate them? How do you think about in the next generation of model, what is worth revisiting? Like, yeah, they're just kind of like, you know, you worked on so many ideas that end up being influential, but like in the moment, they might not feel that way necessarily. Yeah.Jeff Dean [00:03:52]: I mean, I think distillation was originally motivated because we were seeing that we had a very large image data set at the time, you know, 300 million images that we could train on. And we were seeing that if you create specialists for different subsets of those image categories, you know, this one's going to be really good at sort of mammals, and this one's going to be really good at sort of indoor room scenes or whatever, and you can cluster those categories and train on an enriched stream of data after you do pre-training on a much broader set of images. You get much better performance. If you then treat that whole set of maybe 50 models you've trained as a large ensemble, but that's not a very practical thing to serve, right? So distillation really came about from the idea of, okay, what if we want to actually serve that and train all these independent sort of expert models and then squish it into something that actually fits in a form factor that you can actually serve? And that's, you know, not that different from what we're doing today. You know, often today we're instead of having an ensemble of 50 models. We're having a much larger scale model that we then distill into a much smaller scale model.Shawn Wang [00:05:09]: Yeah. A part of me also wonders if distillation also has a story with the RL revolution. So let me maybe try to articulate what I mean by that, which is you can, RL basically spikes models in a certain part of the distribution. And then you have to sort of, well, you can spike models, but usually sometimes... It might be lossy in other areas and it's kind of like an uneven technique, but you can probably distill it back and you can, I think that the sort of general dream is to be able to advance capabilities without regressing on anything else. And I think like that, that whole capability merging without loss, I feel like it's like, you know, some part of that should be a distillation process, but I can't quite articulate it. I haven't seen much papers about it.Jeff Dean [00:06:01]: Yeah, I mean, I tend to think of one of the key advantages of distillation is that you can have a much smaller model and you can have a very large, you know, training data set and you can get utility out of making many passes over that data set because you're now getting the logits from the much larger model in order to sort of coax the right behavior out of the smaller model that you wouldn't otherwise get with just the hard labels. And so, you know, I think that's what we've observed. Is you can get, you know, very close to your largest model performance with distillation approaches. And that seems to be, you know, a nice sweet spot for a lot of people because it enables us to kind of, for multiple Gemini generations now, we've been able to make the sort of flash version of the next generation as good or even substantially better than the previous generations pro. And I think we're going to keep trying to do that because that seems like a good trend to follow.Shawn Wang [00:07:02]: So, Dara asked, so it was the original map was Flash Pro and Ultra. Are you just sitting on Ultra and distilling from that? Is that like the mother load?Jeff Dean [00:07:12]: I mean, we have a lot of different kinds of models. Some are internal ones that are not necessarily meant to be released or served. Some are, you know, our pro scale model and we can distill from that as well into our Flash scale model. So I think, you know, it's an important set of capabilities to have and also inference time scaling. It can also be a useful thing to improve the capabilities of the model.Shawn Wang [00:07:35]: And yeah, yeah, cool. Yeah. And obviously, I think the economy of Flash is what led to the total dominance. I think the latest number is like 50 trillion tokens. I don't know. I mean, obviously, it's changing every day.Jeff Dean [00:07:46]: Yeah, yeah. But, you know, by market share, hopefully up.Shawn Wang [00:07:50]: No, I mean, there's no I mean, there's just the economics wise, like because Flash is so economical, like you can use it for everything. Like it's in Gmail now. It's in YouTube. Like it's yeah. It's in everything.Jeff Dean [00:08:02]: We're using it more in our search products of various AI mode reviews.Shawn Wang [00:08:05]: Oh, my God. Flash past the AI mode. Oh, my God. Yeah, that's yeah, I didn't even think about that.Jeff Dean [00:08:10]: I mean, I think one of the things that is quite nice about the Flash model is not only is it more affordable, it's also a lower latency. And I think latency is actually a pretty important characteristic for these models because we're going to want models to do much more complicated things that are going to involve, you know, generating many more tokens from when you ask the model to do so. So, you know, if you're going to ask the model to do something until it actually finishes what you ask it to do, because you're going to ask now, not just write me a for loop, but like write me a whole software package to do X or Y or Z. And so having low latency systems that can do that seems really important. And Flash is one direction, one way of doing that. You know, obviously our hardware platforms enable a bunch of interesting aspects of our, you know, serving stack as well, like TPUs, the interconnect between. Chips on the TPUs is actually quite, quite high performance and quite amenable to, for example, long context kind of attention operations, you know, having sparse models with lots of experts. These kinds of things really, really matter a lot in terms of how do you make them servable at scale.Alessio Fanelli [00:09:19]: Yeah. Does it feel like there's some breaking point for like the proto Flash distillation, kind of like one generation delayed? I almost think about almost like the capability as a. In certain tasks, like the pro model today is a saturated, some sort of task. So next generation, that same task will be saturated at the Flash price point. And I think for most of the things that people use models for at some point, the Flash model in two generation will be able to do basically everything. And how do you make it economical to like keep pushing the pro frontier when a lot of the population will be okay with the Flash model? I'm curious how you think about that.Jeff Dean [00:09:59]: I mean, I think that's true. If your distribution of what people are asking people, the models to do is stationary, right? But I think what often happens is as the models become more capable, people ask them to do more, right? So, I mean, I think this happens in my own usage. Like I used to try our models a year ago for some sort of coding task, and it was okay at some simpler things, but wouldn't do work very well for more complicated things. And since then, we've improved dramatically on the more complicated coding tasks. And now I'll ask it to do much more complicated things. And I think that's true, not just of coding, but of, you know, now, you know, can you analyze all the, you know, renewable energy deployments in the world and give me a report on solar panel deployment or whatever. That's a very complicated, you know, more complicated task than people would have asked a year ago. And so you are going to want more capable models to push the frontier in the absence of what people ask the models to do. And that also then gives us. Insight into, okay, where does the, where do things break down? How can we improve the model in these, these particular areas, uh, in order to sort of, um, make the next generation even better.Alessio Fanelli [00:11:11]: Yeah. Are there any benchmarks or like test sets they use internally? Because it's almost like the same benchmarks get reported every time. And it's like, all right, it's like 99 instead of 97. Like, how do you have to keep pushing the team internally to it? Or like, this is what we're building towards. Yeah.Jeff Dean [00:11:26]: I mean, I think. Benchmarks, particularly external ones that are publicly available. Have their utility, but they often kind of have a lifespan of utility where they're introduced and maybe they're quite hard for current models. You know, I, I like to think of the best kinds of benchmarks are ones where the initial scores are like 10 to 20 or 30%, maybe, but not higher. And then you can sort of work on improving that capability for, uh, whatever it is, the benchmark is trying to assess and get it up to like 80, 90%, whatever. I, I think once it hits kind of 95% or something, you get very diminishing returns from really focusing on that benchmark, cuz it's sort of, it's either the case that you've now achieved that capability, or there's also the issue of leakage in public data or very related kind of data being, being in your training data. Um, so we have a bunch of held out internal benchmarks that we really look at where we know that wasn't represented in the training data at all. There are capabilities that we want the model to have. Um, yeah. Yeah. Um, that it doesn't have now, and then we can work on, you know, assessing, you know, how do we make the model better at these kinds of things? Is it, we need different kind of data to train on that's more specialized for this particular kind of task. Do we need, um, you know, a bunch of, uh, you know, architectural improvements or some sort of, uh, model capability improvements, you know, what would help make that better?Shawn Wang [00:12:53]: Is there, is there such an example that you, uh, a benchmark inspired in architectural improvement? Like, uh, I'm just kind of. Jumping on that because you just.Jeff Dean [00:13:02]: Uh, I mean, I think some of the long context capability of the, of the Gemini models that came, I guess, first in 1.5 really were about looking at, okay, we want to have, um, you know,Shawn Wang [00:13:15]: immediately everyone jumped to like completely green charts of like, everyone had, I was like, how did everyone crack this at the same time? Right. Yeah. Yeah.Jeff Dean [00:13:23]: I mean, I think, um, and once you're set, I mean, as you say that needed single needle and a half. Hey, stack benchmark is really saturated for at least context links up to 1, 2 and K or something. Don't actually have, you know, much larger than 1, 2 and 8 K these days or two or something. We're trying to push the frontier of 1 million or 2 million context, which is good because I think there are a lot of use cases where. Yeah. You know, putting a thousand pages of text or putting, you know, multiple hour long videos and the context and then actually being able to make use of that as useful. Try to, to explore the über graduation are fairly large. But the single needle in a haystack benchmark is sort of saturated. So you really want more complicated, sort of multi-needle or more realistic, take all this content and produce this kind of answer from a long context that sort of better assesses what it is people really want to do with long context. Which is not just, you know, can you tell me the product number for this particular thing?Shawn Wang [00:14:31]: Yeah, it's retrieval. It's retrieval within machine learning. It's interesting because I think the more meta level I'm trying to operate at here is you have a benchmark. You're like, okay, I see the architectural thing I need to do in order to go fix that. But should you do it? Because sometimes that's an inductive bias, basically. It's what Jason Wei, who used to work at Google, would say. Exactly the kind of thing. Yeah, you're going to win. Short term. Longer term, I don't know if that's going to scale. You might have to undo that.Jeff Dean [00:15:01]: I mean, I like to sort of not focus on exactly what solution we're going to derive, but what capability would you want? And I think we're very convinced that, you know, long context is useful, but it's way too short today. Right? Like, I think what you would really want is, can I attend to the internet while I answer my question? Right? But that's not going to happen. I think that's going to be solved by purely scaling the existing solutions, which are quadratic. So a million tokens kind of pushes what you can do. You're not going to do that to a trillion tokens, let alone, you know, a billion tokens, let alone a trillion. But I think if you could give the illusion that you can attend to trillions of tokens, that would be amazing. You'd find all kinds of uses for that. You would have attend to the internet. You could attend to the pixels of YouTube and the sort of deeper representations that we can find. You could attend to the form for a single video, but across many videos, you know, on a personal Gemini level, you could attend to all of your personal state with your permission. So like your emails, your photos, your docs, your plane tickets you have. I think that would be really, really useful. And the question is, how do you get algorithmic improvements and system level improvements that get you to something where you actually can attend to trillions of tokens? Right. In a meaningful way. Yeah.Shawn Wang [00:16:26]: But by the way, I think I did some math and it's like, if you spoke all day, every day for eight hours a day, you only generate a maximum of like a hundred K tokens, which like very comfortably fits.Jeff Dean [00:16:38]: Right. But if you then say, okay, I want to be able to understand everything people are putting on videos.Shawn Wang [00:16:46]: Well, also, I think that the classic example is you start going beyond language into like proteins and whatever else is extremely information dense. Yeah. Yeah.Jeff Dean [00:16:55]: I mean, I think one of the things about Gemini's multimodal aspects is we've always wanted it to be multimodal from the start. And so, you know, that sometimes to people means text and images and video sort of human-like and audio, audio, human-like modalities. But I think it's also really useful to have Gemini know about non-human modalities. Yeah. Like LIDAR sensor data from. Yes. Say, Waymo vehicles or. Like robots or, you know, various kinds of health modalities, x-rays and MRIs and imaging and genomics information. And I think there's probably hundreds of modalities of data where you'd like the model to be able to at least be exposed to the fact that this is an interesting modality and has certain meaning in the world. Where even if you haven't trained on all the LIDAR data or MRI data, you could have, because maybe that's not, you know, it doesn't make sense in terms of trade-offs of. You know, what you include in your main pre-training data mix, at least including a little bit of it is actually quite useful. Yeah. Because it sort of tempts the model that this is a thing.Shawn Wang [00:18:04]: Yeah. Do you believe, I mean, since we're on this topic and something I just get to ask you all the questions I always wanted to ask, which is fantastic. Like, are there some king modalities, like modalities that supersede all the other modalities? So a simple example was Vision can, on a pixel level, encode text. And DeepSeq had this DeepSeq CR paper that did that. Vision. And Vision has also been shown to maybe incorporate audio because you can do audio spectrograms and that's, that's also like a Vision capable thing. Like, so, so maybe Vision is just the king modality and like. Yeah.Jeff Dean [00:18:36]: I mean, Vision and Motion are quite important things, right? Motion. Well, like video as opposed to static images, because I mean, there's a reason evolution has evolved eyes like 23 independent ways, because it's such a useful capability for sensing the world around you, which is really what we want these models to be. So I think the only thing that we can be able to do is interpret the things we're seeing or the things we're paying attention to and then help us in using that information to do things. Yeah.Shawn Wang [00:19:05]: I think motion, you know, I still want to shout out, I think Gemini, still the only native video understanding model that's out there. So I use it for YouTube all the time. Nice.Jeff Dean [00:19:15]: Yeah. Yeah. I mean, it's actually, I think people kind of are not necessarily aware of what the Gemini models can actually do. Yeah. Like I have an example I've used in one of my talks. It had like, it was like a YouTube highlight video of 18 memorable sports moments across the last 20 years or something. So it has like Michael Jordan hitting some jump shot at the end of the finals and, you know, some soccer goals and things like that. And you can literally just give it the video and say, can you please make me a table of what all these different events are? What when the date is when they happened? And a short description. And so you get like now an 18 row table of that information extracted from the video, which is, you know, not something most people think of as like a turn video into sequel like table.Alessio Fanelli [00:20:11]: Has there been any discussion inside of Google of like, you mentioned tending to the whole internet, right? Google, it's almost built because a human cannot tend to the whole internet and you need some sort of ranking to find what you need. Yep. That ranking is like much different for an LLM because you can expect a person to look at maybe the first five, six links in a Google search versus for an LLM. Should you expect to have 20 links that are highly relevant? Like how do you internally figure out, you know, how do we build the AI mode that is like maybe like much broader search and span versus like the more human one? Yeah.Jeff Dean [00:20:47]: I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. I mean, I think even pre-language model based work, you know, our ranking systems would be built to start. With a giant number of web pages in our index, many of them are not relevant. So you identify a subset of them that are relevant with very lightweight kinds of methods. You know, you're down to like 30,000 documents or something. And then you gradually refine that to apply more and more sophisticated algorithms and more and more sophisticated sort of signals of various kinds in order to get down to ultimately what you show, which is, you know, the final 10 results or, you know, 10 results plus. Other kinds of information. And I think an LLM based system is not going to be that dissimilar, right? You're going to attend to trillions of tokens, but you're going to want to identify, you know, what are the 30,000 ish documents that are with the, you know, maybe 30 million interesting tokens. And then how do you go from that into what are the 117 documents I really should be paying attention to in order to carry out the tasks that the user has asked? And I think, you know, you can imagine systems where you have, you know, a lot of highly parallel processing to identify those initial 30,000 candidates, maybe with very lightweight kinds of models. Then you have some system that sort of helps you narrow down from 30,000 to the 117 with maybe a little bit more sophisticated model or set of models. And then maybe the final model is the thing that looks. So the 117 things that might be your most capable model. So I think it has to, it's going to be some system like that, that is really enables you to give the illusion of attending to trillions of tokens. Sort of the way Google search gives you, you know, not the illusion, but you are searching the internet, but you're finding, you know, a very small subset of things that are, that are relevant.Shawn Wang [00:22:47]: Yeah. I often tell a lot of people that are not steeped in like Google search history that, well, you know, like Bert was. Like he was like basically immediately inside of Google search and that improves results a lot, right? Like I don't, I don't have any numbers off the top of my head, but like, I'm sure you guys, that's obviously the most important numbers to Google. Yeah.Jeff Dean [00:23:08]: I mean, I think going to an LLM based representation of text and words and so on enables you to get out of the explicit hard notion of, of particular words having to be on the page, but really getting at the notion of this topic of this page or this page. Paragraph is highly relevant to this query. Yeah.Shawn Wang [00:23:28]: I don't think people understand how much LLMs have taken over all these very high traffic system, very high traffic. Yeah. Like it's Google, it's YouTube. YouTube has this like semantics ID thing where it's just like every token or every item in the vocab is a YouTube video or something that predicts the video using a code book, which is absurd to me for YouTube size.Jeff Dean [00:23:50]: And then most recently GROK also for, for XAI, which is like, yeah. I mean, I'll call out even before LLMs were used extensively in search, we put a lot of emphasis on softening the notion of what the user actually entered into the query.Shawn Wang [00:24:06]: So do you have like a history of like, what's the progression? Oh yeah.Jeff Dean [00:24:09]: I mean, I actually gave a talk in, uh, I guess, uh, web search and data mining conference in 2009, uh, where we never actually published any papers about the origins of Google search, uh, sort of, but we went through sort of four or five or six. generations, four or five or six generations of, uh, redesigning of the search and retrieval system, uh, from about 1999 through 2004 or five. And that talk is really about that evolution. And one of the things that really happened in 2001 was we were sort of working to scale the system in multiple dimensions. So one is we wanted to make our index bigger, so we could retrieve from a larger index, which always helps your quality in general. Uh, because if you don't have the page in your index, you're going to not do well. Um, and then we also needed to scale our capacity because we were, our traffic was growing quite extensively. Um, and so we had, you know, a sharded system where you have more and more shards as the index grows, you have like 30 shards. And then if you want to double the index size, you make 60 shards so that you can bound the latency by which you respond for any particular user query. Um, and then as traffic grows, you add, you add more and more replicas of each of those. And so we eventually did the math that realized that in a data center where we had say 60 shards and, um, you know, 20 copies of each shard, we now had 1200 machines, uh, with disks. And we did the math and we're like, Hey, one copy of that index would actually fit in memory across 1200 machines. So in 2001, we introduced, uh, we put our entire index in memory and what that enabled from a quality perspective was amazing. Um, and so we had more and more replicas of each of those. Before you had to be really careful about, you know, how many different terms you looked at for a query, because every one of them would involve a disk seek on every one of the 60 shards. And so you, as you make your index bigger, that becomes even more inefficient. But once you have the whole index in memory, it's totally fine to have 50 terms you throw into the query from the user's original three or four word query, because now you can add synonyms like restaurant and restaurants and cafe and, uh, you know, things like that. Uh, bistro and all these things. And you can suddenly start, uh, sort of really, uh, getting at the meaning of the word as opposed to the exact semantic form the user typed in. And that was, you know, 2001, very much pre LLM, but really it was about softening the, the strict definition of what the user typed in order to get at the meaning.Alessio Fanelli [00:26:47]: What are like principles that you use to like design the systems, especially when you have, I mean, in 2001, the internet is like. Doubling, tripling every year in size is not like, uh, you know, and I think today you kind of see that with LLMs too, where like every year the jumps in size and like capabilities are just so big. Are there just any, you know, principles that you use to like, think about this? Yeah.Jeff Dean [00:27:08]: I mean, I think, uh, you know, first, whenever you're designing a system, you want to understand what are the sort of design parameters that are going to be most important in designing that, you know? So, you know, how many queries per second do you need to handle? How big is the internet? How big is the index you need to handle? How much data do you need to keep for every document in the index? How are you going to look at it when you retrieve things? Um, what happens if traffic were to double or triple, you know, will that system work well? And I think a good design principle is you're going to want to design a system so that the most important characteristics could scale by like factors of five or 10, but probably not beyond that because often what happens is if you design a system for X. And something suddenly becomes a hundred X, that would enable a very different point in the design space that would not make sense at X. But all of a sudden at a hundred X makes total sense. So like going from a disk space index to a in memory index makes a lot of sense once you have enough traffic, because now you have enough replicas of the sort of state on disk that those machines now actually can hold, uh, you know, a full copy of the, uh, index and memory. Yeah. And that all of a sudden enabled. A completely different design that wouldn't have been practical before. Yeah. Um, so I'm, I'm a big fan of thinking through designs in your head, just kind of playing with the design space a little before you actually do a lot of writing of code. But, you know, as you said, in the early days of Google, we were growing the index, uh, quite extensively. We were growing the update rate of the index. So the update rate actually is the parameter that changed the most. Surprising. So it used to be once a month.Shawn Wang [00:28:55]: Yeah.Jeff Dean [00:28:56]: And then we went to a system that could update any particular page in like sub one minute. Okay.Shawn Wang [00:29:02]: Yeah. Because this is a competitive advantage, right?Jeff Dean [00:29:04]: Because all of a sudden news related queries, you know, if you're, if you've got last month's news index, it's not actually that useful for.Shawn Wang [00:29:11]: News is a special beast. Was there any, like you could have split it onto a separate system.Jeff Dean [00:29:15]: Well, we did. We launched a Google news product, but you also want news related queries that people type into the main index to also be sort of updated.Shawn Wang [00:29:23]: So, yeah, it's interesting. And then you have to like classify whether the page is, you have to decide which pages should be updated and what frequency. Oh yeah.Jeff Dean [00:29:30]: There's a whole like, uh, system behind the scenes that's trying to decide update rates and importance of the pages. So even if the update rate seems low, you might still want to recrawl important pages quite often because, uh, the likelihood they change might be low, but the value of having updated is high.Shawn Wang [00:29:50]: Yeah, yeah, yeah, yeah. Uh, well, you know, yeah. This, uh, you know, mention of latency and, and saving things to this reminds me of one of your classics, which I have to bring up, which is latency numbers. Every programmer should know, uh, was there a, was it just a, just a general story behind that? Did you like just write it down?Jeff Dean [00:30:06]: I mean, this has like sort of eight or 10 different kinds of metrics that are like, how long does a cache mistake? How long does branch mispredict take? How long does a reference domain memory take? How long does it take to send, you know, a packet from the U S to the Netherlands or something? Um,Shawn Wang [00:30:21]: why Netherlands, by the way, or is it, is that because of Chrome?Jeff Dean [00:30:25]: Uh, we had a data center in the Netherlands, um, so, I mean, I think this gets to the point of being able to do the back of the envelope calculations. So these are sort of the raw ingredients of those, and you can use them to say, okay, well, if I need to design a system to do image search and thumb nailing or something of the result page, you know, how, what I do that I could pre-compute the image thumbnails. I could like. Try to thumbnail them on the fly from the larger images. What would that do? How much dis bandwidth than I need? How many des seeks would I do? Um, and you can sort of actually do thought experiments in, you know, 30 seconds or a minute with the sort of, uh, basic, uh, basic numbers at your fingertips. Uh, and then as you sort of build software using higher level libraries, you kind of want to develop the same intuitions for how long does it take to, you know, look up something in this particular kind of.Shawn Wang [00:31:21]: I'll see you next time.Shawn Wang [00:31:51]: Which is a simple byte conversion. That's nothing interesting. I wonder if you have any, if you were to update your...Jeff Dean [00:31:58]: I mean, I think it's really good to think about calculations you're doing in a model, either for training or inference.Jeff Dean [00:32:09]: Often a good way to view that is how much state will you need to bring in from memory, either like on-chip SRAM or HBM from the accelerator. Attached memory or DRAM or over the network. And then how expensive is that data motion relative to the cost of, say, an actual multiply in the matrix multiply unit? And that cost is actually really, really low, right? Because it's order, depending on your precision, I think it's like sub one picodule.Shawn Wang [00:32:50]: Oh, okay. You measure it by energy. Yeah. Yeah.Jeff Dean [00:32:52]: Yeah. I mean, it's all going to be about energy and how do you make the most energy efficient system. And then moving data from the SRAM on the other side of the chip, not even off the off chip, but on the other side of the same chip can be, you know, a thousand picodules. Oh, yeah. And so all of a sudden, this is why your accelerators require batching. Because if you move, like, say, the parameter of a model from SRAM on the, on the chip into the multiplier unit, that's going to cost you a thousand picodules. So you better make use of that, that thing that you moved many, many times with. So that's where the batch dimension comes in. Because all of a sudden, you know, if you have a batch of 256 or something, that's not so bad. But if you have a batch of one, that's really not good.Shawn Wang [00:33:40]: Yeah. Yeah. Right.Jeff Dean [00:33:41]: Because then you paid a thousand picodules in order to do your one picodule multiply.Shawn Wang [00:33:46]: I have never heard an energy-based analysis of batching.Jeff Dean [00:33:50]: Yeah. I mean, that's why people batch. Yeah. Ideally, you'd like to use batch size one because the latency would be great.Shawn Wang [00:33:56]: The best latency.Jeff Dean [00:33:56]: But the energy cost and the compute cost inefficiency that you get is quite large. So, yeah.Shawn Wang [00:34:04]: Is there a similar trick like, like, like you did with, you know, putting everything in memory? Like, you know, I think obviously NVIDIA has caused a lot of waves with betting very hard on SRAM with Grok. I wonder if, like, that's something that you already saw with, with the TPUs, right? Like that, that you had to. Uh, to serve at your scale, uh, you probably sort of saw that coming. Like what, what, what hardware, uh, innovations or insights were formed because of what you're seeing there?Jeff Dean [00:34:33]: Yeah. I mean, I think, you know, TPUs have this nice, uh, sort of regular structure of 2D or 3D meshes with a bunch of chips connected. Yeah. And each one of those has HBM attached. Um, I think for serving some kinds of models, uh, you know, you, you pay a lot higher cost. Uh, and time latency, um, bringing things in from HBM than you do bringing them in from, uh, SRAM on the chip. So if you have a small enough model, you can actually do model parallelism, spread it out over lots of chips and you actually get quite good throughput improvements and latency improvements from doing that. And so you're now sort of striping your smallish scale model over say 16 or 64 chips. Uh, but as if you do that and it all fits in. In SRAM, uh, that can be a big win. So yeah, that's not a surprise, but it is a good technique.Alessio Fanelli [00:35:27]: Yeah. What about the TPU design? Like how much do you decide where the improvements have to go? So like, this is like a good example of like, is there a way to bring the thousand picojoules down to 50? Like, is it worth designing a new chip to do that? The extreme is like when people say, oh, you should burn the model on the ASIC and that's kind of like the most extreme thing. How much of it? Is it worth doing an hardware when things change so quickly? Like what was the internal discussion? Yeah.Jeff Dean [00:35:57]: I mean, we, we have a lot of interaction between say the TPU chip design architecture team and the sort of higher level modeling, uh, experts, because you really want to take advantage of being able to co-design what should future TPUs look like based on where we think the sort of ML research puck is going, uh, in some sense, because, uh, you know, as a hardware designer for ML and in particular, you're trying to design a chip starting today and that design might take two years before it even lands in a data center. And then it has to sort of be a reasonable lifetime of the chip to take you three, four or five years. So you're trying to predict two to six years out where, what ML computations will people want to run two to six years out in a very fast changing field. And so having people with interest. Interesting ML research ideas of things we think will start to work in that timeframe or will be more important in that timeframe, uh, really enables us to then get, you know, interesting hardware features put into, you know, TPU N plus two, where TPU N is what we have today.Shawn Wang [00:37:10]: Oh, the cycle time is plus two.Jeff Dean [00:37:12]: Roughly. Wow. Because, uh, I mean, sometimes you can squeeze some changes into N plus one, but, you know, bigger changes are going to require the chip. Yeah. Design be earlier in its lifetime design process. Um, so whenever we can do that, it's generally good. And sometimes you can put in speculative features that maybe won't cost you much chip area, but if it works out, it would make something, you know, 10 times as fast. And if it doesn't work out, well, you burned a little bit of tiny amount of your chip area on that thing, but it's not that big a deal. Uh, sometimes it's a very big change and we want to be pretty sure this is going to work out. So we'll do like lots of carefulness. Uh, ML experimentation to show us, uh, this is actually the, the way we want to go. Yeah.Alessio Fanelli [00:37:58]: Is there a reverse of like, we already committed to this chip design so we can not take the model architecture that way because it doesn't quite fit?Jeff Dean [00:38:06]: Yeah. I mean, you, you definitely have things where you're going to adapt what the model architecture looks like so that they're efficient on the chips that you're going to have for both training and inference of that, of that, uh, generation of model. So I think it kind of goes both ways. Um, you know, sometimes you can take advantage of, you know, lower precision things that are coming in a future generation. So you can, might train it at that lower precision, even if the current generation doesn't quite do that. Mm.Shawn Wang [00:38:40]: Yeah. How low can we go in precision?Jeff Dean [00:38:43]: Because people are saying like ternary is like, uh, yeah, I mean, I'm a big fan of very low precision because I think that gets, that saves you a tremendous amount of time. Right. Because it's picojoules per bit that you're transferring and reducing the number of bits is a really good way to, to reduce that. Um, you know, I think people have gotten a lot of luck, uh, mileage out of having very low bit precision things, but then having scaling factors that apply to a whole bunch of, uh, those, those weights. Scaling. How does it, how does it, okay.Shawn Wang [00:39:15]: Interesting. You, so low, low precision, but scaled up weights. Yeah. Huh. Yeah. Never considered that. Yeah. Interesting. Uh, w w while we're on this topic, you know, I think there's a lot of, um, uh, this, the concept of precision at all is weird when we're sampling, you know, uh, we just, at the end of this, we're going to have all these like chips that I'll do like very good math. And then we're just going to throw a random number generator at the start. So, I mean, there's a movement towards, uh, energy based, uh, models and processors. I'm just curious if you've, obviously you've thought about it, but like, what's your commentary?Jeff Dean [00:39:50]: Yeah. I mean, I think. There's a bunch of interesting trends though. Energy based models is one, you know, diffusion based models, which don't sort of sequentially decode tokens is another, um, you know, speculative decoding is a way that you can get sort of an equivalent, very small.Shawn Wang [00:40:06]: Draft.Jeff Dean [00:40:07]: Batch factor, uh, for like you predict eight tokens out and that enables you to sort of increase the effective batch size of what you're doing by a factor of eight, even, and then you maybe accept five or six of those tokens. So you get. A five, a five X improvement in the amortization of moving weights, uh, into the multipliers to do the prediction for the, the tokens. So these are all really good techniques and I think it's really good to look at them from the lens of, uh, energy, real energy, not energy based models, um, and, and also latency and throughput, right? If you look at things from that lens, that sort of guides you to. Two solutions that are gonna be, uh, you know, better from, uh, you know, being able to serve larger models or, you know, equivalent size models more cheaply and with lower latency.Shawn Wang [00:41:03]: Yeah. Well, I think, I think I, um, it's appealing intellectually, uh, haven't seen it like really hit the mainstream, but, um, I do think that, uh, there's some poetry in the sense that, uh, you know, we don't have to do, uh, a lot of shenanigans if like we fundamentally. Design it into the hardware. Yeah, yeah.Jeff Dean [00:41:23]: I mean, I think there's still a, there's also sort of the more exotic things like analog based, uh, uh, computing substrates as opposed to digital ones. Uh, I'm, you know, I think those are super interesting cause they can be potentially low power. Uh, but I think you often end up wanting to interface that with digital systems and you end up losing a lot of the power advantages in the digital to analog and analog to digital conversions. You end up doing, uh, at the sort of boundaries. And periphery of that system. Um, I still think there's a tremendous distance we can go from where we are today in terms of energy efficiency with sort of, uh, much better and specialized hardware for the models we care about.Shawn Wang [00:42:05]: Yeah.Alessio Fanelli [00:42:06]: Um, any other interesting research ideas that you've seen, or like maybe things that you cannot pursue a Google that you would be interested in seeing researchers take a step at, I guess you have a lot of researchers. Yeah, I guess you have enough, but our, our research.Jeff Dean [00:42:21]: Our research portfolio is pretty broad. I would say, um, I mean, I think, uh, in terms of research directions, there's a whole bunch of, uh, you know, open problems and how do you make these models reliable and able to do much longer, kind of, uh, more complex tasks that have lots of subtasks. How do you orchestrate, you know, maybe one model that's using other models as tools in order to sort of build, uh, things that can accomplish, uh, you know, much more. Yeah. Significant pieces of work, uh, collectively, then you would ask a single model to do. Um, so that's super interesting. How do you get more verifiable, uh, you know, how do you get RL to work for non-verifiable domains? I think it's a pretty interesting open problem because I think that would broaden out the capabilities of the models, the improvements that you're seeing in both math and coding. Uh, if we could apply those to other less verifiable domains, because we've come up with RL techniques that actually enable us to do that. Uh, effectively, that would, that would really make the models improve quite a lot. I think.Alessio Fanelli [00:43:26]: I'm curious, like when we had Noam Brown on the podcast, he said, um, they already proved you can do it with deep research. Um, you kind of have it with AI mode in a way it's not verifiable. I'm curious if there's any thread that you think is interesting there. Like what is it? Both are like information retrieval of JSON. So I wonder if it's like the retrieval is like the verifiable part. That you can score or what are like, yeah, yeah. How, how would you model that, that problem?Jeff Dean [00:43:55]: Yeah. I mean, I think there are ways of having other models that can evaluate the results of what a first model did, maybe even retrieving. Can you have another model that says, is this things, are these things you retrieved relevant? Or can you rate these 2000 things you retrieved to assess which ones are the 50 most relevant or something? Um, I think those kinds of techniques are actually quite effective. Sometimes I can even be the same model, just prompted differently to be a, you know, a critic as opposed to a, uh, actual retrieval system. Yeah.Shawn Wang [00:44:28]: Um, I do think like there, there is that, that weird cliff where like, it feels like we've done the easy stuff and then now it's, but it always feels like that every year. It's like, oh, like we know, we know, and the next part is super hard and nobody's figured it out. And, uh, exactly with this RLVR thing where like everyone's talking about, well, okay, how do we. the next stage of the non-verifiable stuff. And everyone's like, I don't know, you know, Ellen judge.Jeff Dean [00:44:56]: I mean, I feel like the nice thing about this field is there's lots and lots of smart people thinking about creative solutions to some of the problems that we all see. Uh, because I think everyone sort of sees that the models, you know, are great at some things and they fall down around the edges of those things and, and are not as capable as we'd like in those areas. And then coming up with good techniques and trying those. And seeing which ones actually make a difference is sort of what the whole research aspect of this field is, is pushing forward. And I think that's why it's super interesting. You know, if you think about two years ago, we were struggling with GSM, eight K problems, right? Like, you know, Fred has two rabbits. He gets three more rabbits. How many rabbits does he have? That's a pretty far cry from the kinds of mathematics that the models can, and now you're doing IMO and Erdos problems in pure language. Yeah. Yeah. Pure language. So that is a really, really amazing jump in capabilities in, you know, in a year and a half or something. And I think, um, for other areas, it'd be great if we could make that kind of leap. Uh, and you know, we don't exactly see how to do it for some, some areas, but we do see it for some other areas and we're going to work hard on making that better. Yeah.Shawn Wang [00:46:13]: Yeah.Alessio Fanelli [00:46:14]: Like YouTube thumbnail generation. That would be very helpful. We need that. That would be AGI. We need that.Shawn Wang [00:46:20]: That would be. As far as content creators go.Jeff Dean [00:46:22]: I guess I'm not a YouTube creator, so I don't care that much about that problem, but I guess, uh, many people do.Shawn Wang [00:46:27]: It does. Yeah. It doesn't, it doesn't matter. People do judge books by their covers as it turns out. Um, uh, just to draw a bit on the IMO goal. Um, I'm still not over the fact that a year ago we had alpha proof and alpha geometry and all those things. And then this year we were like, screw that we'll just chuck it into Gemini. Yeah. What's your reflection? Like, I think this, this question about. Like the merger of like symbolic systems and like, and, and LMS, uh, was a very much core belief. And then somewhere along the line, people would just said, Nope, we'll just all do it in the LLM.Jeff Dean [00:47:02]: Yeah. I mean, I think it makes a lot of sense to me because, you know, humans manipulate symbols, but we probably don't have like a symbolic representation in our heads. Right. We have some distributed representation that is neural net, like in some way of lots of different neurons. And activation patterns firing when we see certain things and that enables us to reason and plan and, you know, do chains of thought and, you know, roll them back now that, that approach for solving the problem doesn't seem like it's going to work. I'm going to try this one. And, you know, in a lot of ways we're emulating what we intuitively think, uh, is happening inside real brains in neural net based models. So it never made sense to me to have like completely separate. Uh, discrete, uh, symbolic things, and then a completely different way of, of, uh, you know, thinking about those things.Shawn Wang [00:47:59]: Interesting. Yeah. Uh, I mean, it's maybe seems obvious to you, but it wasn't obvious to me a year ago. Yeah.Jeff Dean [00:48:06]: I mean, I do think like that IMO with, you know, translating to lean and using lean and then the next year and also a specialized geometry model. And then this year switching to a single unified model. That is roughly the production model with a little bit more inference budget, uh, is actually, you know, quite good because it shows you that the capabilities of that general model have improved dramatically and, and now you don't need the specialized model. This is actually sort of very similar to the 2013 to 16 era of machine learning, right? Like it used to be, people would train separate models for lots of different, each different problem, right? I have, I want to recognize street signs and something. So I train a street sign. Recognition recognition model, or I want to, you know, decode speech recognition. I have a speech model, right? I think now the era of unified models that do everything is really upon us. And the question is how well do those models generalize to new things they've never been asked to do and they're getting better and better.Shawn Wang [00:49:10]: And you don't need domain experts. Like one of my, uh, so I interviewed ETA who was on, who was on that team. Uh, and he was like, yeah, I, I don't know how they work. I don't know where the IMO competition was held. I don't know the rules of it. I just trained the models, the training models. Yeah. Yeah. And it's kind of interesting that like people with these, this like universal skill set of just like machine learning, you just give them data and give them enough compute and they can kind of tackle any task, which is the bitter lesson, I guess. I don't know. Yeah.Jeff Dean [00:49:39]: I mean, I think, uh, general models, uh, will win out over specialized ones in most cases.Shawn Wang [00:49:45]: Uh, so I want to push there a bit. I think there's one hole here, which is like, uh. There's this concept of like, uh, maybe capacity of a model, like abstractly a model can only contain the number of bits that it has. And, uh, and so it, you know, God knows like Gemini pro is like one to 10 trillion parameters. We don't know, but, uh, the Gemma models, for example, right? Like a lot of people want like the open source local models that are like that, that, that, and, and, uh, they have some knowledge, which is not necessary, right? Like they can't know everything like, like you have the. The luxury of you have the big model and big model should be able to capable of everything. But like when, when you're distilling and you're going down to the small models, you know, you're actually memorizing things that are not useful. Yeah. And so like, how do we, I guess, do we want to extract that? Can we, can we divorce knowledge from reasoning, you know?Jeff Dean [00:50:38]: Yeah. I mean, I think you do want the model to be most effective at reasoning if it can retrieve things, right? Because having the model devote precious parameter space. To remembering obscure facts that could be looked up is actually not the best use of that parameter space, right? Like you might prefer something that is more generally useful in more settings than this obscure fact that it has. Um, so I think that's always attention at the same time. You also don't want your model to be kind of completely detached from, you know, knowing stuff about the world, right? Like it's probably useful to know how long the golden gate be. Bridges just as a general sense of like how long are bridges, right? And, uh, it should have that kind of knowledge. It maybe doesn't need to know how long some teeny little bridge in some other more obscure part of the world is, but, uh, it does help it to have a fair bit of world knowledge and the bigger your model is, the more you can have. Uh, but I do think combining retrieval with sort of reasoning and making the model really good at doing multiple stages of retrieval. Yeah.Shawn Wang [00:51:49]: And reasoning through the intermediate retrieval results is going to be a, a pretty effective way of making the model seem much more capable, because if you think about, say, a personal Gemini, yeah, right?Jeff Dean [00:52:01]: Like we're not going to train Gemini on my email. Probably we'd rather have a single model that, uh, we can then use and use being able to retrieve from my email as a tool and have the model reason about it and retrieve from my photos or whatever, uh, and then make use of that and have multiple. Um, you know, uh, stages of interaction. that makes sense.Alessio Fanelli [00:52:24]: Do you think the vertical models are like, uh, interesting pursuit? Like when people are like, oh, we're building the best healthcare LLM, we're building the best law LLM, are those kind of like short-term stopgaps or?Jeff Dean [00:52:37]: No, I mean, I think, I think vertical models are interesting. Like you want them to start from a pretty good base model, but then you can sort of, uh, sort of viewing them, view them as enriching the data. Data distribution for that particular vertical domain for healthcare, say, um, we're probably not going to train or for say robotics. We're probably not going to train Gemini on all possible robotics data. We, you could train it on because we want it to have a balanced set of capabilities. Um, so we'll expose it to some robotics data, but if you're trying to build a really, really good robotics model, you're going to want to start with that and then train it on more robotics data. And then maybe that would. It's multilingual translation capability, but improve its robotics capabilities. And we're always making these kind of, uh, you know, trade-offs in the data mix that we train the base Gemini models on. You know, we'd love to include data from 200 more languages and as much data as we have for those languages, but that's going to displace some other capabilities of the model. It won't be as good at, um, you know, Pearl programming, you know, it'll still be good at Python programming. Cause we'll include it. Enough. Of that, but there's other long tail computer languages or coding capabilities that it may suffer on or multi, uh, multimodal reasoning capabilities may suffer. Cause we didn't get to expose it to as much data there, but it's really good at multilingual things. So I, I think some combination of specialized models, maybe more modular models. So it'd be nice to have the capability to have those 200 languages, plus this awesome robotics model, plus this awesome healthcare, uh, module that all can be knitted together to work in concert and called upon in different circumstances. Right? Like if I have a health related thing, then it should enable using this health module in conjunction with the main base model to be even better at those kinds of things. Yeah.Shawn Wang [00:54:36]: Installable knowledge. Yeah.Jeff Dean [00:54:37]: Right.Shawn Wang [00:54:38]: Just download as a, as a package.Jeff Dean [00:54:39]: And some of that installable stuff can come from retrieval, but some of it probably should come from preloaded training on, you know, uh, a hundred billion tokens or a trillion tokens of health data. Yeah.Shawn Wang [00:54:51]: And for listeners, I think, uh, I will highlight the Gemma three end paper where they, there was a little bit of that, I think. Yeah.Alessio Fanelli [00:54:56]: Yeah. I guess the question is like, how many billions of tokens do you need to outpace the frontier model improvements? You know, it's like, if I have to make this model better healthcare and the main. Gemini model is still improving. Do I need 50 billion tokens? Can I do it with a hundred, if I need a trillion healthcare tokens, it's like, they're probably not out there that you don't have, you know, I think that's really like the.Jeff Dean [00:55:21]: Well, I mean, I think healthcare is a particularly challenging domain, so there's a lot of healthcare data that, you know, we don't have access to appropriately, but there's a lot of, you know, uh, healthcare organizations that want to train models on their own data. That is not public healthcare data, uh, not public health. But public healthcare data. Um, so I think there are opportunities there to say, partner with a large healthcare organization and train models for their use that are going to be, you know, more bespoke, but probably, uh, might be better than a general model trained on say, public data. Yeah.Shawn Wang [00:55:58]: Yeah. I, I believe, uh, by the way, also this is like somewhat related to the language conversation. Uh, I think one of your, your favorite examples was you can put a low resource language in the context and it just learns. Yeah.Jeff Dean [00:56:09]: Oh, yeah, I think the example we used was Calamon, which is truly low resource because it's only spoken by, I think 120 people in the world and there's no written text.Shawn Wang [00:56:20]: So, yeah. So you can just do it that way. Just put it in the context. Yeah. Yeah. But I think your whole data set in the context, right.Jeff Dean [00:56:27]: If you, if you take a language like, uh, you know, Somali or something, there is a fair bit of Somali text in the world that, uh, or Ethiopian Amharic or something, um, you know, we probably. Yeah. Are not putting all the data from those languages into the Gemini based training. We put some of it, but if you put more of it, you'll improve the capabilities of those models.Shawn Wang [00:56:49]: Yeah.Jeff Dean [00:56:49]:

From a sequence starting in 2025. You can join, live, each Tuesday, 7.30 p.m. Ireland time (the same as UK time)! Information about the sequence can be found here: https://first164.blogspot.com/p/zoom164.html

On November 4, 2025, the Roman Catholic Dicastery for the Doctrine of the Faith issued a document titled Mater Populi Fidelis (MPF, Latin for “Mother of the Faithful People”). Signed by Pope Leo XIV, its primary purpose is indicated in Paragraph 22: Roman Catholics are no longer to refer to Mary as the Co-Redemptrix.This is a step in the right direction, but the title “Mother of the Faithful” alone indicates that this document is not Biblical. The Bible only knows one mother of all believers: “…Jerusalem which is above…is the mother of us all” (Galatians 4:26). Additionally, MPF favorably cites dozens of documents from the Magisterium that allege dozens of unbiblical doctrines regarding Mary. The meaning of MPF is also still a matter of debate within the Catholic Church.

Send us a textEver notice how quickly we blame our bodies for our feelings? We dig into A Course in Miracles' bold reminder that mind is cause and the body is a neutral device for learning and communication. By returning cause to thought, we dissolve the exhausting loop where symptoms, circumstances, and biographies get treated as the source of fear. That shift doesn't deny care; it frees care from panic and restores gentle responsibility: honest noticing, clear choosing, and willingness to let correction happen.From there, we tackle a subtle trap: awe. When admiration becomes awe among equals, hierarchy sneaks in. Pedestals feel inspiring at first, then quietly breed dependency, over-caring, and fear of loss. We share simple tells—monitoring someone's state, feeling responsible for their peace, or shrinking around their perceived power—and offer a clean replacement: recognition. Equal worth, shared mind, steady respect, genuine gratitude. This clears space for guidance to lead without specialness distorting the bond.We also re-anchor awe where it belongs: with the Creator. Properly placed, awe doesn't make us small; it confirms safety, source, and real power. Misplaced, it confuses awesome with fearful and keeps us braced against the very love that heals. Through stories, practical examples, and ACIM references, we show how to stop using the body as a defense, how to meet sensations as mirrors instead of meanings, and how to let miracles arise when defenses fall away. The result is less rumination, fewer bodily narratives, and more ease in everyday choices.If this conversation sparked clarity, subscribe, share with a friend who loves ACIM, and leave a review with your biggest takeaway. Your notes help others find the show and join the practice.Support the show

From a sequence starting in 2025. You can join, live, each Tuesday, 7.30 p.m. Ireland time (the same as UK time)! Information about the sequence can be found here: https://first164.blogspot.com/p/zoom164.html

Title: The Greatest Paragraph Ever Written Preacher: H B Charles, Jr. Series: Words From Old Friends Passage: Romans 3:21-26

Defending the proclamation without contention