Last week, @FactoryAI announced a $50M Series B round and launched Droids, its new generation of development agents. The AI coding agent space is already crowded. Claude Code is one of the more popular ones, but there are plenty of other really good solutions, like Amp, Codex, https://t.co/TWxZuEz8Rr
— Sid Bharath (@Siddharth87) Sep 30, 2025
Sorceress summons a lava elemental for battle. Created with @Kling_ai 2.5 turbo Image to video. 🔊🔊🎧 https://t.co/gpFbJ2YpNJ
— Stevie Mac (@StevieMac03) Sep 30, 2025
There’s a new way to access data, and Uber is teaching us how to do it. They built Finch. An AI agent that lives in Slack. Understands business language. Writes SQL. Delivers answers in seconds. In my latest article, I break down: - How Finch actually works - The trade-offs https://t.co/xzLl1lDC4h
— Raul Junco (@RaulJuncoV) Sep 30, 2025
Build full stack AI apps with no API keys. All Google models are FREE for a week in Lovable AI. We've been building so many nano banana powered apps internally, and we think you should too :) https://t.co/Nn283niwVX
— Alexandre Pesant (@AlexandrePesant) Sep 29, 2025
Lovable cloud just dropped. Here's what it means: • Add API's, Databases or even Stripe directly into your build in Lovable • One-click OpenAI API • Baked in MCP's • Custom backends No more work around, just unlimited possibilities! https://t.co/ov5L6YawNL
— damien (@damienghader) Sep 29, 2025
Claude Code Camp is back—and this time, it’s for the non-coders. Nearly 500 people joined our last one. This edition is for writers, founders, marketers, and operators who want to use Claude like a pro without touching code. Learn workflows, prompts, and templates to 10x your https://t.co/ejNinvrn4f
— Every 📧 (@every) Sep 29, 2025
Coding models go through chaotic eras (no clear best model, tradeoffs everywhere) and stable eras (a X.5 Sonnet model is the best in the world) Glad to once again enter an era of stability. Welcome to the world, Sonnet 4.5. https://t.co/8QTcrw6bSn
— Thariq (@trq212) Sep 29, 2025
I just listened to an AI generate a 90-minute conversation and I honestly can't tell you which parts were real anymore 🤯 Microsoft Research released VibeVoice three days ago and I've been testing it nonstop. This isn't like the other TTS models. Those sound fine for 30 seconds, https://t.co/hs0w0HdVTv
— Robert Youssef (@rryssf_) Sep 29, 2025
when software is cheap to build, ecosystems are the edge https://t.co/DEZEu862oD
— Dan Shipper 📧 (@danshipper) Sep 29, 2025
font pairing formula! https://t.co/4SLGxh8iVx
— Pietro Baudin (@pietrobaudin) Sep 29, 2025
🔗 Alternatives to Using Pure Black (#000000) for Text and Backgrounds https://t.co/ab1bSCEH4I
— UI/UX Savior (@UiSavior) Sep 28, 2025
> Models will autonomously work for 8 hour days by mid-2026. > At least one model will match the performance of human experts across many industries before the end of 2026. > By the end of 2027, models will frequently outperform experts on many tasks. https://t.co/3F5a4SsIdM
— Smoke-away (@SmokeAwayyy) Sep 28, 2025
Japanese colours you can use instead of #ffffff for your UI design. https://t.co/LLzeFE1Ion
— UI/UX Savior (@UiSavior) Sep 28, 2025
so @FactoryAI built world class agentic harnesses on top of existing models (prob diff for each model) and it works great, there’s tons of alpha in building harnesses for your task, it’s a big part of the future of agent dev (with and w/o further RL training) if you’re also https://t.co/j5rckDfDJR
— Viv (@Vtrivedy10) Sep 28, 2025
BREAKING: OpenAI just released Prompt Packs for every job. 300+ ready-to-use prompts for: → IT → Sales → Product → Managers → Engineers → Marketing → Executives → Customer Success https://t.co/iXXcjVR9iQ
— Gina Acosta (@ginacostag_) Sep 28, 2025
You're in middle of a MLE interview at Spotify. "Our real-time recommendation model for the Discover Weekly playlist has high inference costs. You propose adding a cache to reduce latency and cost. What is the single biggest risk of this approach?" You: "Cache invalidation is
— Ashutosh Maheshwari (@asmah2107) Sep 28, 2025
Stumbled on this gem last night. Hands down the best PMF framework I’ve come across. At Slice we’ve been intuitively working through this with some success but little structure. This nails it, making execution easier and giving the team a clear way to asses our depth of PMF. https://t.co/OIDMjl4aPA
— Farouk Ismail (@FaroukIsmail_) Sep 27, 2025
Sam Altman literally leaked how to become a billionaire https://t.co/qD8lXzEw01
— Yu Lin (@basicprompts) Sep 27, 2025
Less talking, more building: 𝗠𝘂𝗹𝘁𝗶-𝗔𝗴𝗲𝗻𝘁 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗦𝘆𝘀𝘁𝗲𝗺 𝗳𝗿𝗼𝗺 𝘀𝗰𝗿𝗮𝘁𝗰𝗵! 👇 Some weeks ago I released an episode of my Newsletter and an update to the 𝗔𝗜 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝘀 𝗛𝗮𝗻𝗱𝗯𝗼𝗼𝗸 GitHub repository. There I implemented a Deep https://t.co/mjXW9FPtt2
— Aurimas Griciūnas (@Aurimas_Gr) Sep 27, 2025
Huge effort from the team, proud to be a part of it 📈 https://t.co/ZgVCYpkx41
— Abhay Singhal (@_AbhaySinghal) Sep 26, 2025
Chidorigafuchi Park, Tokyo, Japan. 🌸 https://t.co/lalSbW6j4t
— JAPAN 🇯🇵 (@japanvistas) Sep 26, 2025
very important work on a new eval https://t.co/CWAZGnNAkG
— Sam Altman (@sama) Sep 25, 2025
new piece by me on @every today! https://t.co/CAI5JshySa
— Dan Shipper 📧 (@danshipper) Sep 26, 2025
Today we’re introducing GDPval, a new evaluation that measures AI on real-world, economically valuable tasks. Evals ground progress in evidence instead of speculation and help track how AI improves at the kind of work that matters most. https://t.co/uKPPDldVNS
— OpenAI (@OpenAI) Sep 25, 2025
Roon (@tszzl) on Thinking Machines and their recent research: “New labs are incentivized to publish their work openly, it’s a great recruiting mechanism.” Facebook once followed the same playbook, open-sourcing everything to attract top talent. https://t.co/V4D8iKubdH https://t.co/GZxMMgD2HT
— TBPN (@tbpn) Sep 26, 2025
Pod Up. Pure Jensen. Listen twice at 1.5x - lots of nuance in the words that matter. 🇺🇸 https://t.co/RrJCDTVx6p
— Brad Gerstner (@altcap) Sep 26, 2025
Life hack: There are actually two websites that allow you to legally opt out of virtually all physical junk mail. They look awful but that’s because they hope you won’t use them. They work. (Links in reply.) https://t.co/4M5QcMuBbx https://t.co/QRC9haLIkI
— Austen Allred (@Austen) Sep 25, 2025
BREAKING: Factory Raises $50M Series B from NEA, Sequoia Capital, NVIDIA, JPMorgan & Abstract Ventures The droids are here. @FactoryAI CEO & Co-Founder Matan Grinberg (@matanSF) joins me to talk.. - Droids. - Codegen margins - Strategic investors - Why agent-native is a step https://t.co/1gmmS6h69R https://t.co/qLmKLWE9rs
— Molly O’Shea (@MollySOShea) Sep 25, 2025
From 0 to $15M ARR in 90 days of launch - Emergent Labs is India’s fastest growing AI startup🤯 Emergent was founded a year ago by Dunzo’s former CTO Mukund Jha - they now have 57 team members (75% of headcount) in India The velocity of product shipping & pace of ARR addition
— Rahul Mathur (@Rahul_J_Mathur) Sep 25, 2025
@alex_avoigt What matters is that the cycle time of the overall production process can get to ~10 seconds and there is a path to cars coming off the line every 5 seconds. I posted about Tesla taking this new path based on physics first principles long ago. Btw, anyone is welcome to copy
— Elon Musk (@elonmusk) Sep 25, 2025
Wan Animate isn’t just a character swap. It generates clothing + hair physics that move with you. All you do is: 🎥 Upload a video 🖼️ Add a reference image 💬 Drop a prompt Try it out here 👇 https://t.co/kkHSiIYbEh
— GLIF (@heyglif) Sep 23, 2025
More about time. About time! https://t.co/zF7Y9Vrniq
— David Deutsch (@DavidDeutschOxf) Sep 23, 2025
Abundant Intelligence: https://t.co/sVOhzBeE3A
— Sam Altman (@sama) Sep 23, 2025
This will be epic! @genspark_ai is taking the Agent world by storm! You will learn a ton from @sang_wen https://t.co/t86i59Iny0
— Michael Neuvirth (@LetsGoCure) Sep 23, 2025
$=[],t=0,draw=_=>t++?background(0,14)+($=[...$.slice(-5e3),...[...'#つぶやきProcessing'].map(p5.Vector.random3D)]).map(_=>stroke(255,96).point((_.x+=sin(r=((_.x*2*(k=_.x*4+2^_.y*4|2+_.z*4)^_.y*k+t/299)&1)*2-1)/99*_.z)*109+270,(_.y+=cos(r)/99*_.z)*109+270)):createCanvas(540,540) https://t.co/QXUfVgJPqj
— ア (@yuruyurau) Nov 1, 2023
Peter Thiel on what you should learn from Steve Jobs: https://t.co/fKILlmRTRO
— David Senra (@FoundersPodcast) Sep 22, 2025
NEW Sim AI DESTROYS N8N? (FREE!) 🤯 Want the full guide? DM me. https://t.co/Ll01ng2ktZ
— Julian Goldie SEO (@JulianGoldieSEO) Sep 22, 2025
mem_brane: Inside an artificial mind (installation simulation) https://t.co/ME3a18ArgX
— 𝕭𝖏ø𝖗𝖓 𝕾𝖙𝖆𝖆𝖑 (@_nonfigurativ_) Sep 22, 2025
Origami space engineering https://t.co/VcpJWwA6pk
— Science girl (@gunsnrosesgirl3) Sep 21, 2025
Okay… Nano Banana isn’t just for fun edits. It can make your photos look like they belong in Fortune 500 covers, Vogue spreads, or even Hollywood posters. Here are 10 pro prompts [Bookmark this for later 👇] https://t.co/lgufeXsckb
— Devin AI News and Tools (@toolandtea) Sep 21, 2025
$1,000,000,000+ raised... How can a guy from a tiny town in Illinois, with no engineering degree, be the one building the most powerful humanoid company in the world? The story makes no sense. Until it does. [Long Thread 🧵] https://t.co/ble9d811Gy
— Ilir Aliu - eu/acc (@IlirAliu_) Sep 21, 2025
I just came across a paper on a method called "AutoRefine," and it's making me rethink how we should build AI reasoning systems. Someone figured out a surprisingly simple way to make retrieval-augmented LLMs significantly more accurate, and it's not about a bigger model. For https://t.co/2hiAMzMDet
— Carlos E. Perez (@IntuitMachine) Sep 21, 2025
I’ve read The Making of Prince of Persia cover to cover many times. It’s a remarkable account of how technical constraints can catalyze one of the most fun game designs. It’s a field guide to turning limits into taste. @jmechner's notes on tools, playtesting, and https://t.co/bXj5DsJVZe
— Leo ⌛ (@Leozayaat) Sep 21, 2025
I love how he has articulated what he wants clearly. But people still don't seem to understand what this is. It's not RAG, it's not notebooklm, it's not lmstudio. It's not a solved problem at all, contrary to what anyone claims. I have also always wanted to something like this, https://t.co/By7PVj3Lpz
— Mahesh Sathiamoorthy (@madiator) Sep 20, 2025
𝒇𝒍𝒖𝒊𝒅𝒊𝒕𝒚 〰️ https://t.co/62oBv1uOpg
— ░ perfectloop ░ (@PERFECTL00P) Sep 20, 2025
“Most older, successful people will tell you that their life was best when they lived it unapologetically, on their own terms.” - @naval https://t.co/FUnHwR2ZCo
— Reads with Ravi (@readswithravi) Sep 20, 2025
Sam Altman: “Hire for values first, aptitude second, and skills third.” Most executive recruiters, Sam argues, will reverse this order: “If you say I need to hire a CFO, they will bring you people with 20 years of experience as a CFO at vaguely similar companies. But https://t.co/gvFT04uyNC
— Startup Archive (@StartupArchive_) Sep 20, 2025
This automation stack keeps my business at 90+% margins 1) Build the automation with Claude Code 2) Set it up using n8n 3) Self Host on Hostinger This is how you integrate your workflows without breaking the bank https://t.co/RBRPxNSBjb
— The Startup Ideas Podcast (SIP) 🧃 (@startupideaspod) Sep 19, 2025
Zuck makes a fantastic point here about building AI/ML teams. Adding more cooks to the AI model kitchen does not speed up cook time. Often when I meet with customers it blows their mind that our ML forecast work across all of Microsoft Finance is done by a single person, me. https://t.co/eCX9XoYbOL
— Michael Tokic (@mftokic) Sep 19, 2025
📁 Matthew McConaughey says he wants a private LLM, fed only with his books, notes, journals, and aspirations, so he can ask it questions and get answers based solely on that information, without any outside influence. https://t.co/ZazO4i5HQY
— Jon Hernandez (@JonhernandezIA) Sep 19, 2025
Lovable for Marketing Ops is here. Surprising fact: small meaningful changes in the first time form experience can result in much better first time completion. Now an AI agent can help you build the perfect first handshake for your next customer. https://t.co/ojhf0dUZm1
— Garry Tan (@garrytan) Sep 18, 2025
I just read a paper from Bloomberg's AI team with valuable nuggets. It’s the first data-driven map of how GenAI is really transforming finance, and the story is not what you think. Everyone says the AI race is simple: bigger models win. We think progress is just a straight
— Carlos E. Perez (@IntuitMachine) Sep 19, 2025
I think I might have discovered the best 2D game engine for vibe coding games. For the past few weeks, I have been making a lovecraftian fishing game with peaceful day time fishing and intense night time combat where you fend off mutated sea creatures. And I've been using https://t.co/p6AgN9Ns1k
— Danny Limanseta (@DannyLimanseta) Sep 19, 2025
I'm ready to accept a definition of "agent" that I think is widely-enough agreed upon to be useful: An LLM agent runs tools in a loop to achieve a goal This is a big piece of personal character development for me! I've been dismissing the term as hopelessly ambiguous for years
— Simon Willison (@simonw) Sep 18, 2025
It's crazy how the greats like Da Vinci & Steve Jobs have carried similar opinions over a difference of nearly 400 years expressed in different contexts. This is just one example but many such incidents around political discourse, economic discourse are a bunch of repeated https://t.co/QPocIkdcyx
— Sairaj (@sairajsk) Sep 18, 2025
Holy smokes! The DeepSeek-R1 team just published the path to superintelligence! “…instead of forcing the model to copy human reasoning steps, it only rewards getting the final answer right, which gives the model freedom to invent its own reasoning strategies that can actually https://t.co/W1AFyBgLAH
— Derya Unutmaz, MD (@DeryaTR_) Sep 18, 2025
Deploy any ML model as an MCP server! LitServe: The easiest way to deploy agents, RAG, and any ML model. Also supports MCP. 100% open-source. https://t.co/WQXdgOYTSo
— Daily Dose of Data Science (@DailyDoseOfDS_) Sep 18, 2025
Working on a waitlist page with cosmic theme in @framer https://t.co/vVoJOtg5Tz
— Skar (@Learning_Skar) Sep 18, 2025
🇨🇳 A parade of transforming military robots from China. multi-terrain spiders (wheeled, flying, amphibious), missile-armed robot dogs, and modular, all-terrain snakes that swim and burrow. https://t.co/PNxMndnqPm
— Rohan Paul (@rohanpaul_ai) Sep 18, 2025
Claude just broke n8n automation forever. This changes everything 🔥 Screenshot any n8n workflow → Get perfect JSON code in seconds → Paste & deploy instantly ✨ P.S. Want the full guide? DM me. https://t.co/kjRs24Gvxb
— Julian Goldie SEO (@JulianGoldieSEO) Sep 18, 2025
this image has singlehandedly saved me from so much overthinking https://t.co/aNMlPGlu0L
— pdawg (@prathamgrv) Sep 16, 2025
Babe wake up: Figma launched vibe designing https://t.co/OuHSqdBDLI
— Satya (@heysatya_) Sep 16, 2025
Anthropic literally just released a full breakdown of which jobs use the most AI https://t.co/BKb10Z4T4L
— Yu Lin (@basicprompts) Sep 16, 2025
Introducing a very special project: The Launch of Rocket Lab. A stunning 300+ page book that takes you behind the scenes of our founding through to today. Our most memorable missions, our toughest challenges, our most nail-biting moments (so far...) all in one place. https://t.co/udNorfpthT
— Rocket Lab (@RocketLab) Sep 16, 2025
Super intriguing idea in this new @GoogleDeepMind paper - shows how to handle the rise of AI agents acting as independent players in the economy. It says that if left unchecked, these agents will create their own economy that connects directly to the human one, which could bring https://t.co/xOYmeNmnCz
— Rohan Paul (@rohanpaul_ai) Sep 16, 2025
One of the best technical blogs I have ever read period! https://t.co/7cO7WSRG9z
— Pramod Goyal (@goyal__pramod) Sep 15, 2025
I’m fascinated with history’s thinker-doers: Cicero, Marcus Aurelius, Benjamin Franklin. There’s no greater achievement than calling your shots: laying out a theoretical worldview, then executing ruthlessly to bend reality. There exist people like this today, who write a https://t.co/N39Kz4Y7AD
— Johnathan Bi (@JohnathanBi) Sep 15, 2025
A Universe of Data Generated with Hailuo 02 @Hailuo_AI It gives me immense pleasure to tweak prompts until I find something new. When I start, I have no idea where I’m going, but I discover it along the way. Final prompt below: https://t.co/cULkuauc9b
— Alex Patrascu (@maxescu) Sep 13, 2025
a=(y,d=mag(k=(y<9?4+sin(y^8)*9:y/4+cos(y))*cos(i),e=y/6-13)+cos(i/1199)/2)=>point((q=y*k/d*(2+sin(d*2+y-t*16))+79)*cos(c=d/6-t+(i&5))+200,q*d/9*sin(c+i%2*7)+200) t=0,draw=$=>{t||createCanvas(w=400,w);background(9).stroke(w,56);for(t+=PI/480,i=2e4;i--;)a(i/980)}#つぶやきProcessing https://t.co/LoTkHCeTND
— ア (@yuruyurau) Sep 13, 2025
How a book written in 1910 can teach you calculus in 30 seconds: https://t.co/39G40WXrVW
— PyQuant News 🐍 (@pyquantnews) Sep 13, 2025
Some advice hits you so hard it rewires how you think https://t.co/vL9zfqtV2N
— Ruben Dominguez Ibar (@rdominguezibar) Sep 12, 2025
the ladder of product value in 1 photo: relevant → functional → desirable → magical https://t.co/9juLI0JlHp
— GREG ISENBERG (@gregisenberg) Sep 13, 2025
Agentic AI Roadmap 2025 https://t.co/YAeCj4frdJ
— Python Programming (@PythonPr) Sep 13, 2025
The colours: The design: https://t.co/nzqgZ9CbAs
— echo° (@echoGFX_) Sep 12, 2025
We just hit the next level of AI lipsync 🔥 This is the breakthrough that will finally allow you to create your own TV shows and IPs. It’s fast. It’s cheap. And it completely changes the viewer experience. The silent era of AI cartoons is over. https://t.co/uB8JXQ78vE
— Framer 🇱🇹 (@0xFramer) Sep 12, 2025
Recent design work for b2b AI companies https://t.co/gHESqRGQoK
— Kyle Anthony Miller (@kyleanthony) Sep 12, 2025
In Chapter 2 of Systems Engineering for Agentic Applications, we build our own MCP. Not another MCP server but our own protocol We'll discover - What is MCP - What does MCP do - Why would we even need MCP https://t.co/dePYglid4L https://t.co/f4j6eIxp2A
— Dominik Tornow (@DominikTornow) Sep 12, 2025
understanding GEO, AI search and how to get traffic (and money) from it: my head exploded when I heard that "AI search converts at 10–40% vs 1–2% from SEO." okay, so what do we need to know about generative search aka GEO: ai search (perplexity, chatgpt, gemini) is still just https://t.co/pG19u2u6DL
— GREG ISENBERG (@gregisenberg) Sep 12, 2025
Never thought I'd get to design mechs but here we are. https://t.co/mLADc6YLKi
— Pastor (@FussyPastor) Sep 10, 2025
Holy shit they’re doing on-policy RL by just deploying the model to prod lmao that’s so baller. also 2 hrs for a training step makes our 10 minute steps feel lightning fast @hamishivi … they probably have a bigger batch size though 😅 https://t.co/LSAV5OgVXU https://t.co/zjjDCysK7z
— Saurabh Shah (@saurabh_shah2) Sep 12, 2025
AI-generated news is here and it's not going away: Patch CEO @warrenstjohn says his company is producing 11,000 local newsletters with zero humans involved. Worth noting: This works in part because *other* publishers are still employing humans to generate news Patch can link to.
— Peter Kafka (@pkafka) Sep 10, 2025
"While FedEx is now a household word, the company that invented overnight delivery was far from an overnight success. The inside story is one of great interest to anyone starting a new business, for it highlights the extraordinary combination of grit and determination, teamwork https://t.co/CCR8tigUAT
— Shane Parrish (@ShaneAParrish) Sep 11, 2025
🔥 Genspark AI Browser now available on Windows and Mac! 💻 On-Device Free AI - The world's first browser letting you choose from 169 AI models to run entirely on-device. No internet required, completely private, lightning fast, and totally free! Beyond traditional browsers - https://t.co/BCKN6ETB78
— Genspark (@genspark_ai) Sep 11, 2025
🧩 Thinking Machines’s mission is to make AI models predictable instead of inconsistent. For example, ask ChatGPT the same question a few times over, and you’re likely to get a wide range of answers. This has largely been accepted in the AI community as a fact — today’s AI https://t.co/Zn4HyYB766 https://t.co/euyw9moU6Q
— Rohan Paul (@rohanpaul_ai) Sep 11, 2025
New course added! We’re launching “Knowledge Graphs for AI Agent API Discovery,” built in collaboration with @SAP and taught by Pavithra G K and Lars Heling. When agents call APIs in the wrong sequence, entire workflows collapse. The challenge? Most agents lack the contextual https://t.co/hIb5ejvmKs
— DeepLearning.AI (@DeepLearningAI) Sep 10, 2025
We built so much into our open source agentic RAG framework, Elysia, that our "bonus features" would be the main selling points of other apps 😅 Seriously - while other frameworks are celebrating their latest chunking strategies or feedback systems, we just casually mention ours https://t.co/3RfWhjoUOx
— Victoria Slocum (@victorialslocum) Sep 10, 2025
Netflix founder Reed Hastings on the tension between performance and loyalty at a startup “The big struggle we had in the early days was the contrast between loyalty and performance,” Reed explains. “Everybody wants loyalty and performance, but they’re at tension.” This is why https://t.co/t9CU0tL9wX
— Startup Archive (@StartupArchive_) Sep 10, 2025
Every serious AI-native company will own + create their own custom agents. Prime Intellect builds the full-stack RL infra to enable this for everyone - from compute, environments to full RL infra - to enable AI-native startups to post-train agents on their own data. https://t.co/ZG3kANrO8Q https://t.co/xyF66YHJmC
— Vincent Weisser (@vincentweisser) Sep 10, 2025
Introducing Tests for ElevenLabs Agents. Run built-in test scenarios to improve the success rate of your agents across: - Tool calling - Human transfers - Complex workflows - Guardrails - Knowledge retrieval https://t.co/Hw2MTkiF57
— ElevenLabs (@elevenlabsio) Sep 9, 2025
How to spot high agency people: They avoid underthinking and overthinking. https://t.co/EttLffPdKk
— George Mack (@george__mack) Sep 9, 2025
Nano Banana + Veo 3 + n8n is insane 🤯 This AI system creates UGC videos at scale using Nano Banana + Veo 3. And runs automatically inside n8n. Perfect for e-comm operators & creative agencies. But instead of manually creating scripts & imagers & videos... Just upload ONE https://t.co/ILlmwrbhZp
— Mike Futia (@mikefutia) Sep 9, 2025
.@HamelHusain and @Sh_reya's online course—AI Evals for Engineers & PMs—is the #1 highest-grossing course on Maven (by far), and consistently brings in large groups from all the major AI labs. This is because they teach something crucial: how to build evaluations that actually https://t.co/r1u1l2bWQu
— Lenny Rachitsky (@lennysan) Sep 9, 2025
with the @cognition fundraise announcement, we're also sharing some new financial/traction details publicly! full thing in replies https://t.co/ogpkdvbu6K
— andrew gao (@itsandrewgao) Sep 8, 2025
Great insights from Bob on FDEs https://t.co/I3FrkoCZ8C
— Chad Wahlquist (@chadwahl) Sep 8, 2025
"True product people have a compulsion to sell directly to consumers, without middlemen fucking things up. Elon was that way." (Full control, over everything, always.) https://t.co/Nc4inrWJHJ
— David Senra (@FoundersPodcast) Sep 8, 2025
Congrats to @cognition for raising over $400M at a $10.2B valuation. Together with @windsurf we have collectively doubled our revenue with our acquisition, and our customers can benefit from the opportunities of both products having synergies between local and cloud agents. We’re
— Jeff Wang (@jeffwsurf) Sep 8, 2025
NVIDIA says SLMs are the future of agents and it makes sense: agents mostly do a small set of repetitive tasks that small language models are already good at. using heavy LLMs for these tasks is like buying a supercomputer for high-school math. you can have 10-30x cost https://t.co/hxzqrZSWyl
— ℏεsam (@Hesamation) Sep 8, 2025
Nano banana + MakeUGC + Veo3 = Ad Factory This agent creates 100s of ads every day - No actors - No product - No $10k/month agencies You're able to clone your comepitors ads with AI Paste there ad -> Pick an avatar and regenerate. Comment "NANO" and I'll send you the https://t.co/2BnG2d1tZS
— George Stock (@georgesttock) Sep 7, 2025
Databricks, Snowflake, and Microsoft are all rapidly trying to bridge the gap between analytics and applications. Palantir's Foundry with Ontology remains the only true e2e platform. CodeStrap's "Code and Connor" Episode 10 releases on Monday, September 8th, 2025, with our https://t.co/TyzqjTOvZR
— CodeStrap (@CodeStrap411) Sep 7, 2025
Superb @AIatMeta paper. 🫡 Speeds up RAG by compressing context into chunk embeddings while keeping answer quality. Up to 30.85x faster first token and up to 16x longer effective context without accuracy drop. RAG prompts paste many retrieved passages, most barely relate, so https://t.co/0wV4riEYGV
— Rohan Paul (@rohanpaul_ai) Sep 7, 2025
Ladies and gentlemen, the generative AI industry: https://t.co/UyGLpnQ2x0
— Luiza Jarovsky, PhD (@LuizaJarovsky) Sep 7, 2025
Another impressive paper by Google DeepMind. It takes a closer look at the limits of embedding-based retrieval. If you work with vector embeddings, bookmark this one. Let's break down the technical details: https://t.co/Orv7QoM1wi
— elvis (@omarsar0) Sep 7, 2025
Everyone is talking about this new OpenAI paper. It's about why LLMs hallucinate. You might want to bookmark this one. Let's break down the technical details: https://t.co/BQ3wSLtiHm
— elvis (@omarsar0) Sep 6, 2025
Awesome way to visualize embeddings powered by Google DeepMind’s Embedding Gemma. Designed for on-device AI, its 308M parameter design enables applications like RAG and semantic search to run directly on your hardware. Demo: huggingface https://t.co/gljxCVjcVn
— Leon (@LeonWohlgemuth) Sep 6, 2025
"Claude Code is not just for coding. It's an AI agent that can do whatever you want it to do." Here's my new tutorial with @AlexFinnX where he showed me exactly how to build a life OS with Claude Code in 25 minutes. He saves 2+ hours daily using: ✅ /researcher to analyze https://t.co/e6tECFVnGW
— Peter Yang (@petergyang) Sep 7, 2025
Forever 21 is using Ai models that look Ai We’re watching the first aggressive wave of Ai adoption hit fashion! https://t.co/v24dlk4Dtk
— Salma (@Salmaaboukarr) Sep 6, 2025
Codex + deep research agents is next-level stuff. Glad I took the bite with Codex CLI. I thought it was just hype. Honestly, it's crazy good! The Windsurf extension is nice too. I will share some cool stuff I have been testing in Codex next week. Stay tuned!
— elvis (@omarsar0) Sep 6, 2025
this was an amazing read https://t.co/TXGcaFxfrz
— maharshi (@mrsiipa) Sep 6, 2025
Meta Superintelligence Labs just made LLMs handle 16x more context and unlocked up to a 31x speedup. 🤯 Their new REFRAG framework rethinks RAG from the ground up to achieve this, all with zero drop in accuracy. Here's how it works: The core problem with long context is https://t.co/nHGupjZ0NI
— Jackson Atkins (@JacksonAtkinsX) Sep 6, 2025
Most people use Claude Code for coding. But my friend @AlexFinnX uses it to run his life. He's built a "Claude Life" operating system that uses AI agents to: - Do research - Curate AI news - Track business metrics - Analyze brain dumps ...and more I asked him to walk me https://t.co/siS2XivZks
— Peter Yang (@petergyang) Sep 6, 2025
Google DeepMind just released a paper stating what we’ve known for 2 years. Their recent paper, “On the Theoretical Limitations of Embedding-Based Retrieval”, shows how vector embeddings fail to retrieve complex semantic relationships as the number of documents in a dataset
— Daniel Davis (@TrustSpooky) Sep 5, 2025
grep (and lightweight semantic search) are all you need 🤔 When you have a “medium” sized dataset e.g. 1000 ArXiv PDFs, we found that an extremely strong Q&A baseline is just giving agents access to the CLI, along with some tools for fast semantic search using static embeddings. https://t.co/sYr1lWadgh https://t.co/OKcCu4qYZh
— Jerry Liu (@jerryjliu0) Sep 5, 2025
[New] from a16z @speedrun: The Outsider founders are winning the AI B2B race we used to say founder x market fit was everything but in this AI wave, Outsider founders who are AI native are outmaneuvering their industry veteran competitors just look at @TennrOfficial https://t.co/TXRudXkaqE
— Troy Kirwin (@tkexpress11) Sep 5, 2025
We've raised our $38M Series B, led by @BessemerVP, with HubSpot Ventures and @SalesforceVC joining. Over 2,000 companies (including @HubSpot, @clickup, @useapolloio) build on @recallai, the API for meeting recording. https://t.co/6nRvmJExs6
— David Gu (@davidruigu) Sep 4, 2025
Announcing our $13M seed to put AI Agents in the hands of every team (code or no-code). So far, we've worked with top companies like Anthropic, Midjourney, Clay, PostHog and tons more to power their customer-facing AI assistants. Today, we’re launching Inkeep Agents: a platform https://t.co/LnIQZtiRJj
— Nick G (@nickgomezc) Sep 5, 2025
I am once again shocked at how much better image retrieval performance you can get if you embed highly opinionated summaries of an image, a summary that came out of a visual language model, than using CLIP embeddings themselves. If you tell the LLM that the summary is going to be
— jason liu (@jxnlco) Sep 5, 2025
You can’t build AI products like other products, for two reasons: 1. AI products are inherently non-deterministic 2. Every AI product must negotiate a tradeoff between agency and control When companies don’t recognize these differences, their AI products face unexpected https://t.co/5g2OtqqYmp
— Lenny Rachitsky (@lennysan) Sep 5, 2025
Huge congrats to the Exa team, and my former landlords! for more background, you should check out @WilliamBryk's talk (for which he won Best Speaker) and his @latentspacepod which was remarkably well received! https://t.co/wCtp78h6FB https://t.co/jI8m88nxmg
— swyx 🇸🇬 (@swyx) Sep 4, 2025
Both DSPy and (especially) GEPA are currently severely under hyped in the AI context engineering world
— tobi lutke (@tobi) Sep 4, 2025
Exa is the truth - organizing the world's information but for AI agents everywhere is an *extreme* unlock for human knowledge https://t.co/MPurqJT4NG
— Garry Tan (@garrytan) Sep 4, 2025
Huge congrats to @exa_ai who just raised a crazy $85M Series B at a $700M valuation! I sat down with @WilliamBryk to discuss why Exa is building the search engine for AI, not humans. https://t.co/5A7EU5yzfX
— Nicolas Dessaigne (@dessaigne) Sep 3, 2025
Function calling & MCP for LLMs, clearly explained (with visuals):
— Avi Chawla (@_avichawla) Sep 3, 2025
The AI opportunity in US healthcare in one simple chart. https://t.co/0tql1UEuJm
— Vaibhav Domkundwar (@vaibhavbetter) Sep 3, 2025
Just concluded a live stream: How Uber built their AI Agent (Genie) using LangGraph (@LangChainAI), @qdrant_engine, Gemini, Ragas and @streamlit In case you missed it, I got you covered - YouTube: https://t.co/yqgbfBBOuY - GitHub: https://t.co/VQA1BF3RP0 Slides in description https://t.co/nGvydUtASf
— Tarun Jain (@TRJ_0751) Aug 31, 2025
did you know ChatGPT can make circuit boards now? https://t.co/MVr3XmBf3n
— Seve (@seveibar) Aug 31, 2025
Naval Ravikant (@naval) on how to build the team that ships: https://t.co/2yc6YUuYbL
— Z Fellows (@zfellows) Aug 31, 2025
Build and deploy LLM agents just using natural language! AutoAgent is the Fully-Automated & Zero-Code LLM Agent Framework that let's you create and deploy LLM agents using just natural language. 100% Open Source https://t.co/CdRACTD6j3
— Sumanth (@Sumanth_077) Aug 31, 2025
