This is NOT realtime voice. This is not some custom GPT with standard voice that talks for 20 mins in response to a question; can't hear properly over a crappy mic and takes 10s to respond. This is STT, TTS and livekit using a locally running node and python setup. It has https://t.co/PJnl71rRkC
— Jason Kneen (@jasonkneen) Jun 30, 2025
We’re releasing Open Researcher next week 👀 It’s an advanced open source web researcher and extractor that uses @firecrawl_dev’s endpoints as primitives. The agent thinks between tool calls for more sophisticated reasoning while traversing the web. Check it out: https://t.co/dCGJ4BAg6y
— Eric Ciarla (hiring) (@ericciarla) Jun 29, 2025
This is the first book with the dark mode. You should try it. https://t.co/NjMfVR9dbw https://t.co/0bkjXZTvG1
— Andriy Burkov (@burkov) Jun 29, 2025
🚨 BREAKING: You can now clone any music without copyright risk. Mureka AI just launched a tool that lets you upload any reference track and generate a custom instrumental in the same style. Here's how it works (+ wild examples):
— Dr Alex Young ⚡️ (@AlexanderFYoung) Jun 29, 2025
Nice list of existing Deep Research Agents: https://t.co/RzQEkp8sye
— elvis (@omarsar0) Jun 28, 2025
>be microsoft >acquire wizardlm team >do absolutely nothing and riddle them with politics >wizardlm team leaves for tencent in may >wizardlm team delivers SOTA reasoning model in june biggest microsoft fumble ever https://t.co/CEySUecpkC
— Casper Hansen (@casper_hansen_) Jun 29, 2025
This guy is genius… Someone gave Claude 4 Sonnet one mega prompt and now it runs his entire marketing engine: → Research → SEO → Ad copy → Content → Strategy Here’s the exact prompt (steal it):
— William Del Principe (@WillDelPrincipe) Jun 28, 2025
@DavidOndrej1 This is the opinion of one development lab. Anthropic put out another article at the time: https://t.co/QRszlEEKjr
— Conner Pope (@digitalconner) Jun 29, 2025
https://t.co/mhYMTVVFay
— David Ondrej (@DavidOndrej1) Jun 28, 2025
It took me 15 days to compile this list of the most important LLM papers. I didn't want to miss any work that led to where we are today. https://t.co/N4K4PdpKfD https://t.co/4OXdT6otNJ
— Pramod Goyal (@goyal__pramod) Jun 28, 2025
A curated collection of papers exploring the path towards Deep Research Agents. https://t.co/49OezVuiJW
— Carlos E. Perez (@IntuitMachine) Jun 28, 2025
Data point on AI's exponential growth (not new of course) 1) $GOOGL Monthly Processed Tokens Up 50x in 12 months 2) $MSFT Tokens Grew 5x in a Similar Time Period Source: Barclays https://t.co/X9IVPeAWpX
— Ray Wang (@rwang07) Jun 27, 2025
New from Ilya Sutskever He talks about AI recursively building better AI. He argues such systems could compress decades of biomedical RAG into months, erasing many diseases and extending life, yet the same loop could outstrip any safety protocol. Prediction fails because https://t.co/aJE6Gpp2Gp
— Rohan Paul (@rohanpaul_ai) Jun 28, 2025
🚨 BREAKING: Claude just turned into an app store. Anthropic dropped a new feature that lets you build and run AI apps inside Claude no API keys, no deploys, no backend. Just chat, build, ship. Here's everything you need to know: https://t.co/LgAEL1PtsX
— Brendan Jowett (@jowettbrendan) Jun 28, 2025
@jerryjliu0 💯 https://t.co/t0pXHxXoj7
— Eleanor Berger (@intellectronica) Jun 26, 2025
RAG vs Agentic RAG https://t.co/5adhFMrSYR
— Alex Xu (@alexxubyte) Jun 26, 2025
6-26-2025 ($) An Interview with Sierra Founder and CEO Bret Taylor About AI Agents and Tech History Lessons https://t.co/9xe2gY8bbB
— Stratechery (@stratechery) Jun 26, 2025
The convergence of AI demand and energy infrastructure is creating a fascinating market dynamic. We're seeing massive capital influx into an already overheated sector where energy infrastructure costs have surged 3-5 years running (maxed turbine capacity, scarce EPC labor,
— Shanu Mathew (@ShanuMathew93) Jun 26, 2025
BofA) AI Accelerators (GPU, ASIC, XPU) 2 While not all computers (including servers) need accelerators, all AI servers must be equipped with accelerators like GPUs, ASICs, and XPUs to be competitive in AI workloads. Types of Accelerators There are two main types of https://t.co/Z9xFLmjmmS
— Jukan Choi (@Jukanlosreve) Jun 26, 2025
Satya with some big statements on the YC podcast! Paraphrased from 15:30 timestamp: "This field is changing so rapidly. I had not even imagined last year, that we would get this far with RL and with test time compute. And it seems pretty limitless. Pre training worked. All the
— Jamin Ball (@jaminball) Jun 26, 2025
Apple will release, to open-source, a 7B diffusion LLM model! Can't wait! https://t.co/2qZ0oouUh8
— Carlos E. Perez (@IntuitMachine) Jun 27, 2025
This guy literally automated an Instagram account using end-to-end AI Agent and got ~5 million views in just 3 weeks. It scrapes the viral Instagram reels and generate similar ones using Veo 3 fully on autopilot. Let that sink in. https://t.co/kzsxfAMBoY
— Shubham Saboo (@Saboo_Shubham_) Jun 26, 2025
DeepMind just dropped a 106-page paper unveiling AlphaGenome. This single model could completely redefine how we discover disease-causing mutations and drug targets. This is massive. 🧵 https://t.co/lhi4S2cd6V
— Dr. Dominic Ng (@DrDominicNg) Jun 25, 2025
BofA is seeing a shift in AI investment towards AI Agent, which is highly positive for overall AI spend. AI Agent requires much higher inference compute, which also benefits from test time compute scaling. While current AI Agent is not yet at commercial readiness, they will https://t.co/RdVIJaCDyE
— Nomad Semi (@MooreMorrisSemi) Jun 26, 2025
“Steven Spielberg said to me, ‘Oh, you’re the director of 'Godzilla Minus One' (2023). I saw it three times,. I couldn’t believe it, because Spielberg is like a god to me, just for what he’s done for the film industry." - Takashi Yamazaki https://t.co/HAurQJRBSm
— cinesthetic. (@TheCinesthetic) Jun 26, 2025
Wrote an intro to evals for long-context Q&A systems: • How it differs from basic Q&A • What dimensions & metrics to eval on • How to build llm-evaluators • How to build eval datasets • Benchmarks: narratives, technical docs, multi-docs https://t.co/XAzPcG7tvf
— Eugene Yan (@eugeneyan) Jun 25, 2025
Thrilled to introduce AlphaGenome, our new DNA sequence model now available via our AlphaGenome API. Really excited to see how the scientific community uses AlphaGenome’s predictions to understand genome function, drive biological discoveries, develop new treatments, and more... https://t.co/uP8Ur6YjWc
— Demis Hassabis (@demishassabis) Jun 25, 2025
Like I said on $NVDA. Lots of sell side firms updating their TAM for datacenter AI, most upping their estimates for @NVIDIA. https://t.co/ioD63vNo0n
— Ben Bajarin (@BenBajarin) Jun 25, 2025
A RAG engine for deep document understanding! RAGFlow lets you build enterprise-grade RAG workflows on complex docs with well-founded citations. Supports multimodal data understanding, web search, deep research, etc. 100% local & open-source with 55k+ stars! https://t.co/wzt0bx6EPb
— Akshay 🚀 (@akshay_pachaar) Jun 24, 2025
Must watch intro to DSPy just out. Incredibly practical and accessible. https://t.co/2qg7JPxt2H https://t.co/14ePZ4GDd1
— DSPy (@DSPyOSS) Jun 25, 2025
How novel is your research? Meet KnoVo, an LLM tool to measure research novelty by comparing a paper’s contributions to others in its citation network Extracts key dimensions (technique, dataset, etc) Scores novelty across those dimensions Visualizes how ideas evolve over time https://t.co/xLKyNlHvdj
— alphaXiv (@askalphaxiv) Jun 24, 2025
the new hot topic is "context engineering" we think LangGraph is really great for enabling completely custom context engineering - but we want to make it even better see our proposal (s/o @sydneyrunkle) for streamlining context management: https://t.co/AhUHhVywvV https://t.co/BGhBVQBk1S
— Harrison Chase (@hwchase17) Jun 24, 2025
0/🚨 AI companies behave differently from traditional SaaS companies, and founders consistently ask us how AI companies are adapting and breaking out. After talking to hundreds of AI companies over the past few years, we think that there are a few emerging principles for https://t.co/KJ511PjC7t
— Kimberly Tan (@kimberlywtan) Jun 24, 2025
Google DeepMind just dropped Gemini Robotics on-device model > runs locally on robots without requiring an internet connection > model performs at almost at par to cloud-based Gemini model > outperforms other on-device models > devs can control and fine-tune the model using https://t.co/4GXHC2Vt2K https://t.co/pQeqYUfRAM
— NIK (@ns123abc) Jun 24, 2025
Much of what worked in traditional SaaS doesn’t hold true in the AI era. In meeting with hundreds of AI founders building for the enterprise, a few core principles have emerged that are shaping how teams are building and growing in today’s market. - Flashy demos are easy, https://t.co/oRZdlzhaq5
— a16z (@a16z) Jun 24, 2025
Holy sh*t… Anthropic just released a prompt engineering bible. It’s free. It’s insanely well written. And it covers real Claude use cases. Every AI builder should bookmark this. Here’s what’s inside + link: https://t.co/B12V3u72RD
— Alex Hughes (@alxnderhughes) Jun 24, 2025
Brian Armstrong's letter to a Product Manager is actually the perfect blueprint for any people manager. Why? Your team IS your product. Let's apply his 5 principles to leading people: https://t.co/M7lHOTCRfD
— Dave Kline (@dklineii) Jun 23, 2025
Checking In on AI and the Big Five A review of the current state of AI through the lens of the Big Five tech companies. https://t.co/eky20nLrwa
— Stratechery (@stratechery) Jun 23, 2025
Yes: The Prestige Limitless Magazine Dreams (new) Black Swan Jiro dreams of sushi The Aviator Kill Bill I + II Foxcatcher Oppenheimer Social Network The Founder Molly's Game Eyes Wide Shut Burnt Ferrari Fight Club Batman (Nolan) Mad Max series Docs: Last Dance + Defiant Ones https://t.co/z5Zbaz3Klf
— Zach 🏴 (@zachpogrob) Jun 22, 2025
Smooth knowledge distillation between different kinds of VLMs is now possible👇 @NVIDIA and @kaist_ai (it's always so interesting to read their papers) have created a universal method called GenRecal. It uses Recalibrator acting like a translator that adjusts how one model https://t.co/TvQG0ktE8e
— Ksenia Se (@Kseniase_) Jun 22, 2025
The AI prompting underneath Cluely, a recent startup that raised $15m from @a16z, has been exposed. It reveals the unique interaction use case that the business supports. Let's dig deep into this so you can do the same! (I'm sure they also have a VC pitch version of this!) https://t.co/NiR8vOXoFl
— Carlos E. Perez (@IntuitMachine) Jun 22, 2025
I think the most interesting part of this is Brian taking the time to explain to a new PM what great looks like. Straight out of Andy Grove's High Output Management: "A manager generally has two ways to raise the level of individual performance of his subordinates: by https://t.co/dlS4OWshRx
— Michael McGuiness (@mikemcg0) Jun 22, 2025
Sam Altman on the Massive AI Product Gap: 'It's an Exceptional Time to Build' "We're in a really interesting time. We haven't been in one of these for a while. The product overhang relative to what the models are capable of is here. The products that people have figured out to https://t.co/qlBUt4lsHJ
— AI Insights (@plzaccelerate) Jun 22, 2025
This study is being massively misinterpreted. College students who wrote an essay with LLM help engaged less with the essay & thus were less engaged when (a total of 9 people) were asked to do similar work weeks later. LLMs do not rot your brain. Being lazy & not learning does. https://t.co/EInE1VnknW
— Ethan Mollick (@emollick) Jun 20, 2025
Cluely's prompts have been leaked (@a16z values this at $15m)! Here's an analysis! Context Reassertion & Intent Echoing The prompt repeatedly orients the agent to "the current moment," prioritizing what’s visible/active right now (screen, audio, transcript end). Pattern https://t.co/bRiTT4rb0x
— Carlos E. Perez (@IntuitMachine) Jun 21, 2025
Kitty Olympics 🐈🐈⬛ 🤯1.5M views in just 5 hours between my Instagram and TikTok — all thanks to the new model from @Hailuo_AI : the Hailuo 02. The physics are insane. I have to say that not every video came out perfect on the first try, but I still got great results really https://t.co/WafbRFrfXc
— Pablo Prompt (@pabloprompt) Jun 19, 2025
Systems of record are recognizing they cannot “take their survival for granted.” One strategy is to acquire : the rationale Salesforce gives for the Informatica acquisition. Another strategy is more defensive - hampering access to the data within the systems of record (SOR). https://t.co/Iz3lHclEMi
— Tomasz Tunguz (@ttunguz) Jun 20, 2025
Interesting slide from the Coatue deck. Interesting test of a firms commitment/investment in AI. Are you getting more efficient? Do any of your execs argue they could get by with "fewer" headcount? Quite a high-bar test. https://t.co/W2MGkAPZuo
— Bill Gurley (@bgurley) Jun 20, 2025
Our team @a16z has gone deep in AI x edtech, meeting hundreds of startups. We just dropped a podcast with our learnings. What AI app have 50% of U.S. teachers tried? Which colleges are mandating "AI fluency"? And who pays most for edtech? Takeaways, clips, and link 👇 https://t.co/Md27B0GjeP
— Olivia Moore (@omooretweets) Jun 20, 2025
Found more from Charlie https://t.co/pQO3P0LyoG https://t.co/RgWMhiFjpn
— Patrick OShaughnessy (@patrick_oshag) Jun 20, 2025
At Perplexity, we’re excited about agentic shopping—helping users discover their next purchase, compare products, and buy seamlessly across many merchants. We’ve expanded our merchant coverage and improved our ability to fulfill complex shopping needs. Let us know what other https://t.co/ywyxIO5PPq
— Dhruv Bhalla (@DhruvBhalla_) Jun 20, 2025
Howard Marks, one of the OG value investors, uses Perplexity to write his memos. Fanboying so hard rn. What other illustrious Benjamin Graham disciples should we show PPLX to? https://t.co/rqDjJvbQsq
— kwak (@dnlkwk) Jun 20, 2025
Yes! Here’s an article I wrote in 2023 about this: https://t.co/tUCAc9yuuE https://t.co/xpxqWXnMKj
— Dan Shipper 📧 (@danshipper) Jun 19, 2025
He is the best teacher in AI. https://t.co/LyJqu7JE9x
— Robert Scoble (@Scobleizer) Jun 19, 2025
I enjoyed this section in the LLM evals course reader. Btw 35% discount here: https://t.co/3botO05aw6 It suggests three gulfs builders have to navigate when building AI systems. First, the Gulf of Comprehension. Don't assume we know our data; that's a trap. Real-world data is https://t.co/rVvb68OKv9
— Eugene Yan (@eugeneyan) Jun 20, 2025
Some 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 might seem simple from the outside. They are not 👇 In my latest Newsletter episode I describe the 𝗘𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗛𝗶𝗲𝗿𝗮𝗿𝗰𝗵𝘆 𝗼𝗳 𝗡𝗲𝗲𝗱𝘀 framework. You can find the deep dive here: https://t.co/1FdsFXtLHZ On a high https://t.co/HdquaMflhr
— Aurimas Griciūnas (@Aurimas_Gr) Jun 19, 2025
https://t.co/x0BuAOTvPT
— Aish (@aish_caliperce) Jun 19, 2025
.@FactoryAI Co-Founder and CTO @EnoReyes breaks down the three core characteristics that make AI systems agentic and reveals why the future of software development requires a fundamental shift from AI-assisted coding to fully agent-driven workflows. Learn how Factory builds https://t.co/ORWYM2IITy
— LangChain (@LangChainAI) Jun 18, 2025
Introducing the OpenAI Podcast—a series of conversations with the people shaping AI. @sama joins @andrewmayne on the first episode to talk about AGI, (wen) GPT-5, privacy, and what comes next. https://t.co/THQOWn8rDw
— OpenAI (@OpenAI) Jun 18, 2025
Spotify founder Daniel Ek is leading a €600m funding in German defence tech startup Helsing, valuing the four-year-old co at €12bn. A biz making battlefield AI, drones, submarines & robo-fighter pilots is now one of Europe’s most valuable startups https://t.co/OKWkuAScl3
— Tim Bradshaw (@tim) Jun 17, 2025
the man who coined “vibe coding” just gave a masterclass in thinking about LLMs as a new type of software and how to think about building new products in this AI era @karpathy @ycombinator https://t.co/VQPvabvNpU
— Adish Jain ☕️ (@_adishj) Jun 17, 2025
The progress of Gemini over the last year + https://t.co/55q4nnYm3g
— Logan Kilpatrick (@OfficialLoganK) Jun 18, 2025
New episode of Uncapped with @sama. Enjoy 🤗 https://t.co/2IxYt3B4Gm
— Jack Altman (@jaltma) Jun 17, 2025
.@satyanadella: The big question for AI builders - "Where does the model end and where does the product begin?" Great conversation between @Microsoft CEO Satya and @ycombinator's @garrytan today at AI Start-up School. Here are my notes: 1. We're riding the fourth wave of https://t.co/QFh7osr60E
— Addy Osmani (@addyosmani) Jun 17, 2025
Live at ai startup school Andrew Ng on speed in startups: -Concrete ideas buy you speed -Concrete idea is one that is specified in enough detail an engineer can build it -Good concrete ideas often come from subject matter experts - their gut can be very good for making decisions https://t.co/HTLnWB3vRd
— jia (@jia_seed) Jun 17, 2025
In early tests: - Creatives made with https://t.co/3HaKJBNxSy saw up to 3x higher CTR - CPA dropped by 30–50% - Brands ran 10x more tests per week Better creatives = better performance. Period. https://t.co/OiSJbdzuk8
— Md Riyazuddin (@riyazmd774) Jun 16, 2025
Google literally published a 68-page prompting masterclass https://t.co/kVbMOx1IVY
— Aadit Sheth (@aaditsh) Jun 16, 2025
AI just learned to fine-tune itself between questions. MIT introduces SEAL, a framework enabling LLMs to self-edit and update their weights via reinforcement learning, all by itself. LLMs consume whatever data they are given, so they stay frozen after pretraining. SEAL teaches https://t.co/rQVPWoSp2W
— Rohan Paul (@rohanpaul_ai) Jun 16, 2025
These 8 n8n automations can replace your entire marketing team. They handle creative testing, research, copy, reporting - and run 24/7 without asking for a raise. Built for businesses that need to move fast without bloated teams or overpriced tools. I use these daily. You
— J.B. (@VibeMarketer_) Jun 16, 2025
I worry that companies are so focused on wringing efficiency from yesterday's technology (and reducing headcount) that they're missing how today's models could transform their actual competitive advantage. The report: https://t.co/S0XlBXAgdt
— Ethan Mollick (@emollick) Jun 16, 2025
Last Friday, Anthropic released a gem - How to build multi-agent research system. Over the weekends, our team had tried to re-create the same architecture, not exactly 1-1, but close to it. Here's how it works and what we've learnt: https://t.co/6rCKQ1ZfTL
— FlowiseAI (@FlowiseAI) Jun 16, 2025
Terence Tao reporting on the same problem we have seen endlessly in other domains: LLM’s produce output that “looks” correct but is often deeply wrong and even stupid on careful inspection. I know of no domain in which this is NOT the case. And yet people seem surprised over https://t.co/GpR2yhx7ZM
— Gary Marcus (@GaryMarcus) Jun 15, 2025
I enjoyed reading this. Valuable insights include: > "Our experiments also revealed binary scoring to be more effective than granular scales. While a 1–10 scale might seem more precise, binary evaluations provide greater consistency, simplicity, and alignment with human https://t.co/tTBj58cpLJ
— Eugene Yan (@eugeneyan) Jun 15, 2025
Here's my conversation with Terence Tao, one of the greatest mathematicians in history. We talk about the hardest problems in mathematics & physics, and how AI might help us humans to solve them. This conversation was a huge honor for me. I can't quite put it into words, but https://t.co/KpeXs0szj9
— Lex Fridman (@lexfridman) Jun 14, 2025
Mixture of experts in LLMs, clearly explained along with visuals: https://t.co/4lGUTqfPpr https://t.co/ZEMlZ5PSey
— Avi Chawla (@_avichawla) Jun 14, 2025
Claude Opus, coordinating four instances of Sonnet as a team, used about 15 times more tokens than normal. (90% performance boost) Jensen has mentioned similar numbers on stage recently. GPT-5 is rumored to be agentic teams based. The demand for compute will continue to increase. https://t.co/Mf4JvJtzdw https://t.co/Ov2DtDUbBK
— Andrew Curran (@AndrewCurran_) Jun 14, 2025
YC is the YC for hard tech https://t.co/qPiGPtBRBX
— Garry Tan (@garrytan) Jun 14, 2025
AI engineering updates are broken. https://t.co/mtnEezgSTv Stop hiding behind vague adjectives. Quantify or don't mention it. Have clear hypotheses. Present actionable data. Good updates drive decisions. Bad ones waste time.
— jason liu (@jxnlco) Jun 14, 2025
Claude research process diagram https://t.co/3tK1yk8Mdx
— Tom Dörr (@tom_doerr) Jun 14, 2025
🚨BREAKING: The world’s first agentic browser just launched. Fellou is the world’s first agentic browser that performs real actions online based on your prompt. Here’s how it works (and why it’s insane)👇
— Artificial Intelligence (AI) • ChatGPT (@chatgptricks) Jun 14, 2025
thanks @huggingface for working with us on a new intro to mcp course. check it out: https://t.co/tm6ezzGxwq
— Theodora Chu (@chu_onthis) Jun 14, 2025
I liked this podcast from @every explaining how to use AI for everything - including research, planning, and user feedback analysis - not just isolated code writing. Summary👇 https://t.co/KVJIcRO0iO https://t.co/H2H6k0TIYR
— Peter Wildeford 🇺🇸🚀 (@peterwildeford) Jun 14, 2025
Transformer vs. Mixture of Experts in LLMs, clearly explained (with visuals):
— Avi Chawla (@_avichawla) Jun 14, 2025
CES 2025 was insane!!! Here are 10 of the best tech and gadgets you shouldn't miss: https://t.co/Ow3vEmobq1
— The AI Colony (@TheAIColony) Jun 14, 2025
Paper: https://t.co/213rEki4rm
— Ethan Mollick (@emollick) Jun 14, 2025
> Traditional SaaS is done (soon) > commoditized > founders clone successful products & halve the price > a race to the bottom, down to $0 > everything is about to look & work the same: the social mediaschedulers, web builders, calendar apps, etc I think this is what comes next:
— John Rush (@johnrushx) Jun 13, 2025
After ~6 years of building these types of architectures (starting with BERT, eg see Baleen), I think calling these multi-agent systems is a distraction. This is just software. Happens to be AI software. It doesn’t seem so complicated once you internalize it’s just a program. https://t.co/PLit8cBdaq https://t.co/Ocgu9DQmGO
— Omar Khattab (@lateinteraction) Jun 13, 2025
Oracle hit a record market cap of $586B and Larry Ellison owns 41% of the company ($240B), making him 2nd richest person in the world. At the end 2010, Oracle valued at $87B and Ellison owned ~27% of it ($24B). How did his ownership of Oracle increase over that span? Since https://t.co/UzvF04V2O8
— Trung Phan (@TrungTPhan) Jun 13, 2025
🚀 Introducing Manus Playbook Each one is like a mini-app with clear guidance for any scenario: 📊 Business | 🎨 Creative | 📈 Sales & Marketing | 🎓 Education | 🎯 Fun & Life Playbook is your step-by-step roadmap in Manus. https://t.co/woeY36FbsO
— ManusAI (@ManusAI_HQ) Jun 12, 2025
Introducing 10x better Exa Remote MCP! Our MCP have tools like web_search, linkedIn_search, company_researcher_, crawling, and much more! Try now: https://t.co/nXpbTFbDDb https://t.co/irIWo3dW7S
— Ishan Goswami (@TheIshanGoswami) Jun 12, 2025
Everyone talks about these two papers, but what do many people miss about them? 1. The Illusion of Thinking by Apple 2. How much do language models memorize? by FAIR, Google DeepMind, Cornell University, and NVIDIA They describe the same underlying breakdown – a model’s coping https://t.co/BkuYXJHoC7
— TuringPost (@TheTuringPost) Jun 11, 2025
Most AI companies say they're trying to augment workers, not replacing them. But this week, I wrote about @MechanizeWork, a buzzy new AI start-up whose explicit goal is automating all jobs. I don't love that, but I'm glad they're not hiding the ball. https://t.co/2gsb6bBpmO
— Kevin Roose (@kevinroose) Jun 11, 2025
This face was NEVER alive. It’s a damn JPEG that learned to talk. You can generate an image… and make it act, sell, cry, explain, with perfect tone, emotion & timing. Here’s how to pull it off in 30 seconds 👇 https://t.co/Soq9tNd5TD
— EyeingAI (@EyeingAI) Jun 12, 2025
Netflix founder Reed Hastings’ #1 piece of advice for young CEOs When asked for his best advice for young CEOs, Reed responds: “I would say memorize the first 86 pages of Jim Collins’s first big book, Beyond Entrepreneurship. Someone gave it to be in 1994, and I’ve read it https://t.co/Mv8zMgZkTz
— Startup Archive (@StartupArchive_) Jun 12, 2025
OpenAI has open-sourced a demo of a UI testing agent that uses the OpenAI Computer-Using Agent (CUA) model, the Responses API, and Playwright to automate frontend testing https://t.co/r8HUeMGjbw
— Tibor Blaho (@btibor91) Jun 12, 2025
About AI and national security- note the demonstration about five minutes in https://t.co/B8ShU7Yie5
— Eric Schmidt (@ericschmidt) Jun 12, 2025
Pytest for LLM applications! DeepEval is an LLM evaluation framework that makes it easy to test, debug and evaluate LLM applications. 100% open-source and runs locally https://t.co/VwpZ0sR4h9
— Sumanth (@Sumanth_077) Jun 11, 2025
.@pmarca and @jaltma on why AI changes everything. “We’re investing against the thesis that basically all incumbents are going to get nuked, and everything is going to get re-built.” https://t.co/ddpWD2NRfH
— a16z (@a16z) Jun 11, 2025
Frank Slootman: “There’s lots of people in this world who are not really in the arena” Frank comments on why he opens his book with Theodore Roosevelt’s famous “The Man in the Arena” quote: “There’s lots of people in this world who are not really in the arena. They are either https://t.co/zYVvm1mPGz
— Startup Archive (@StartupArchive_) Jun 11, 2025
https://t.co/UH1lbzni41
— Sam Altman (@sama) Jun 10, 2025
Major reasoning models so far with technical reports (focused on those w RL): 2025-01-22 — DeepSeek R1 — https://t.co/FanOPm9oTF 2025-01-22 — Kimi 1.5 — https://t.co/NN8Nr1EAmQ 2025-03-31 — Open-Reasoner-Zero — https://t.co/H5ycSmAkwS 2025-04-10 — Seed 1.5-Thinking —
— Nathan Lambert (@natolambert) Jun 11, 2025
Powerful o3 Pro prompt: 'My company is {company}. Analyze our product, traffic, users, competitors, etc., and find the biggest low-hanging fruit "win" for me that I haven't yet implemented.'
— Matt Shumer (@mattshumer_) Jun 10, 2025
With OpenAI dropping the price of o3 by 80%, today is a great reminder about how important it is to build for where AI is going instead of just what's possible now. You can now get 5X the amount of output today for the same price you were paying yesterday. If you’re building AI
— Aaron Levie (@levie) Jun 11, 2025
I talk about the Ontology as a digital twin of your business processes, encoding the data, logic, and actions then connecting and optimizing your business processes. This view shows off Machinery, a tool used to mine processes embedded in their application. OODA Loop in action https://t.co/xK5TXzY71n
— Chad Wahlquist (@chadwahl) Jun 9, 2025
I recently sat down with the one & only @patrick_oshag to discuss the "State of the Venture Industry 2025." While AI is clearly an exciting new wave, the overall system dynamics are unprecedented, self-reinforcing, & challenging for many participants. https://t.co/aPzjLpTfwl
— Bill Gurley (@bgurley) Jun 10, 2025
Having no experience did not stop me from making a VTOL with world-class range and flight time, all in 90 days. This is the thing I'm most proud of building to date! https://t.co/iUtNeZqFAw
— Tsung Xu (@tsungxu) Jun 9, 2025
🚀 Autonomous Agents That Think, Remember, and Evolve An incredible project by Aaron Brown (from AWS team), showcasing how autonomous agents can move beyond simple tasks to reason, remember, and adapt - powered by Mem0, @awscloud Bedrock, and Strands Agents SDK 💡 From https://t.co/ZbboGMX2PP
— mem0 (@mem0ai) Jun 9, 2025
One of my favorite AI books this year! 🔥 "Building Agentic AI Systems" 🤖 It breaks down agent design in a way that actually makes sense. From reasoning and memory to tool use and multi-step planning. It also provides a comprehensive guide on building LLM-powered agents that https://t.co/vgYtRQK7Y3
— Lucy Wang (@TechWithLucy) Jun 9, 2025
A Personal Update By Vijay Pande Today, I’m sharing that after more than a decade, it’s both exciting and bittersweet that I’ll be stepping away from my role at Andreessen Horowitz. This has been one of the most meaningful chapters of my life, and I’m filled with gratitude as I
— Vijay Pande (@vijaypande) Jun 9, 2025
Training on wrong answers outpaces training on correct ones. 10 times more learning emerges from plausible errors than from truths. Large language models refine their accuracy slowly when they learn only from correct examples. This paper introduces Likra, which trains one https://t.co/iQQ0Q8RQpL
— Rohan Paul (@rohanpaul_ai) Jun 9, 2025
Loved this description of Christopher Nolan: "His curiosity was pure, personal, and delightfully obsessive." https://t.co/KhR0ESlKdr
— David Senra (@FoundersPodcast) May 31, 2025
"I recently rewatched one of Steve Ballmer's speeches at an early Windows conference where he kept repeating the word 'developers.' And he's right, that the companies that get the best developers often win. It's got to be the focus." Mary Meeker https://t.co/74qu7PfBAr
— Tren Griffin (@trengriffin) May 31, 2025
🚨BREAKING: The world’s first open-source AI video tool is here. SkyReels V2 lets you create cinematic content for less than 10% of the cost. It's completely free and open source, with no time limits on videos over 30 seconds. Here’s why this will blow your mind: https://t.co/U8WUAkqPzF
— Artificial Intelligence (Sentient) (@sentient_agency) Jun 8, 2025
Give me a novel to read this week. Want something related to crime, American history, or some masculinity shit (which this is the least masculine thing to say but yeah I want a manly man novel, big deal).
— Sam Parr (@thesamparr) Jun 8, 2025
Nobody gave credit to my prompt.. https://t.co/wR3OY07VxO
— God of Prompt (@godofprompt) Jun 8, 2025
Kyoto, Japan. https://t.co/cNG7xtkVHD
— JAPAN 🇯🇵 (@Japantravelco) Jun 8, 2025
“Arnault had such an appetite for victory and such a capacity for work that he was bound to succeed” https://t.co/nPbvFg5zhn
— David Senra (@FoundersPodcast) Jun 8, 2025
I published this back in 2022 when I was > contributing to Julia and Flux(.)jl at the time > invited by @OfficialLoganK to publish it in JuliaZoid ( @JuliaLanguage 's official medium publication ) those were pre-gsoc days https://t.co/By6aGqEtJK
— Anshuman (e/ia) (@kanpuriyanawab) Jun 7, 2025
https://t.co/mXFBw7X4pN
— Steven Sinofsky (@stevesi) Jun 8, 2025
8 Greatest Laws I know at 35, I wish I knew at 20: – Thread – 1. https://t.co/DuBFim0Xne
— Gentle Man (@Gentle_masc) Jun 7, 2025
The price to become a Generative AI Data Scientist: $12.00 Yet, companies are paying $200,000 for this role. Here's how I did it (and how you can too). https://t.co/ZyKFujHmKC
— 🔥 Matt Dancho (Business Science) 🔥 (@mdancho84) Jun 7, 2025
I read through Mary Meeker's 340-slide Internet Trends report on AI. Pulled the 17 most interesting slides and wrote about them here (puts speed of consumer AI penetration and $1T capex in historical perspective): https://t.co/Bo3Q7jocsH
— Trung Phan (@TrungTPhan) Jun 7, 2025
This should be ILLEGAL… This YT Video Agent is NUTS: It goes from raw idea → full YouTube video → daily uploads without lifting a finger. Here’s what it does: → Pulls fresh video ideas + captions from Google Sheets → Uses GPT-4o to write 10-scene voiceover scripts + https://t.co/6gFSQam8tL
— Damian Player (@damianbplayer) Jun 7, 2025
I've completed the second sweep of my AI Twitter lists, keeping them selective and manageable. I've added a few more. I might split the "AI Community" list into separate categories and have you decide who goes where. Please check them out. If you spot any obvious omissions, let https://t.co/cIUjCIoMKJ https://t.co/k6yUCAfg9j
— Kol Tregaskes (@koltregaskes) Jun 6, 2025
Loved doing a fireside chat w @scottbelsky to talk AI and creativity, in collab w Adobe Some of our topics: - AI enables personalized experiences for everything. Not just personalized entertainment, but also custom software interfaces that look dependent based on the context https://t.co/hl24WQKgsr
— andrew chen (@andrewchen) Jun 6, 2025
Apple just GaryMarcus'd LLM reasoning ability https://t.co/735UMGk4be
— Josh Wolfe (@wolfejosh) Jun 7, 2025
A breakthrough in Fusion energy is closer than most people think. In this Decoded by @ycombinator, we're doing a deep dive on Fusion and what @Helion_Energy is building. https://t.co/s6rhZxm670
— Gustaf Alströmer (@gustaf) Jun 6, 2025
From why Sam Altman is expecting big things in 2026, to how Meta plans to use AI to automate advertising — we round up this week's big stories from the AI revolution https://t.co/bM85oAQwOW https://t.co/CtuzBFDjY1
— Reuters (@Reuters) Jun 4, 2025
Wise beyond his years. A thread of smart things Steve Jobs said when he was 29 years old: https://t.co/JjBnfINPMe
— David Senra (@FoundersPodcast) Jun 5, 2025
2025.23: What Nike Learned About E-Commerce The best Stratechery content from the week of June 2, 2025, including why Nike is working with Amazon, the logic of an Anduril and Meta partnership, and the Japanese rice crisis. https://t.co/9ULzjWpHgN
— Stratechery (@stratechery) Jun 6, 2025
Insane: 90-95% of Claude Code is now written by Claude Code. I asked @mikeyk (Anthropic's CPO) what changes when so much of your code is written by AI? His answer: The bottleneck moves from coding/design to: 1. Upstream: aligning the team around what to build. 2. Downstream: https://t.co/Xq5JX5PJPE https://t.co/aeMHAAx3sT
— Lenny Rachitsky (@lennysan) Jun 5, 2025
one of the best guides on agentic architectures just dropped for free with beautiful in-depth explanations and examples: > patterns in multi-agent architectures > components of ai agents > parallel, network, aggregator, and more we need guides like this. https://t.co/hmgZ2Zi6rG
— ℏεsam (@Hesamation) Jun 5, 2025
In case you missed it, here’s the entire Q&A section from @PalantirTech Annual General Meeting this morning: “We continue to see unrelenting demand for AlP really across a variety of industries.” “One area of particularly explosive growth is the demand we're seeing for warp https://t.co/FHr5kjZ8lh
— Jack Prescott (@JackPrescottX) Jun 5, 2025
One of our top feature requests! I've used this to work with very large knowledge bases in my Projects and it's very handy https://t.co/dmMXj7EElW
— Mike Krieger (@mikeyk) Jun 5, 2025
Introducing Exa Research - agentic search that doesn’t stop until it finds what you need. Research scores 94.9% on SimpleQA, the highest of any research API. We also do it at the lowest price, by using cheaper models for easier tasks, smarter models for harder tasks. The https://t.co/7As1sT80IU
— Exa (@ExaAILabs) Jun 5, 2025
This is a key insight. AI Agents are so well suited for SaaS because the industry has spent 20 years ensuring that there’s an API for nearly every capability that gets shipped. This means AI Agents can instantly become super-users of the system to go and execute work. https://t.co/1QpbYoiOLG
— Aaron Levie (@levie) Jun 4, 2025
I used this n8n workflow to scrape X for top-performing posts in my niche (it's a big part of how I went from 600 → 6k followers in 13 days) most creators scroll for hours to find content ideas that work... whereas I just run this flow and get a full swipe file in minutes - https://t.co/2GTOfmTen3
— J.B. (@VibeMarketer_) Jun 5, 2025
Where is voice AI "working" now? Four areas of the market where I'm seeing the most traction - each of which has different dynamics (and opportunities!) for builders👇 https://t.co/aul7JhVetJ
— Olivia Moore (@omooretweets) Jun 4, 2025
30 lines is all you need to create a MCP Agent using @GoogleDeepMind Gemini 2.5 and remote MCP servers. Here is an example using @cognition_labs Deepwiki MCP server to to chat about repositories. https://t.co/AekXO3SL6P
— Philipp Schmid (@_philschmid) Jun 4, 2025
Meta published a case study of our work on fine tuning llama models for product taxonomy inference https://t.co/E9MNApqVYG
— tobi lutke (@tobi) Jun 4, 2025
Many people underestimate the power of good prompts. Here are 5 frameworks to write effective prompts: [ 🔖 bookmark this to upgrade your prompts ] 1. R–T–F - Act as a (ROLE) - Create a (TASK) - Show as (FORMAT) Example: - ROLE: You’re a B2B SaaS Product Marketer - TASK: https://t.co/2JlrAS1zju
— Andrew Bolis (@AndrewBolis) Jun 3, 2025
$140 billion. That’s what companies spend every year on market research trying to understand their customers. Most of it goes to outdated methods: slow surveys, lagging panels, and static reports. AI is flipping the model. Instead of expensive human surveys, research startups https://t.co/zipnWfTtXi
— a16z (@a16z) Jun 3, 2025
Mary Meeker’s back, and her new AI report has a lot to say about healthcare. Here are 5 key slides: 1. Kaiser is all-in on AI scribes. Since late 2023, over 10,000 physicians and staff have used them to streamline documentation, already logging over 2.5M visits. https://t.co/cUnvzFuL6l
— Healthcare AI Guy (@HealthcareAIGuy) Jun 3, 2025
Last week, @Google dropped a paper on ATLAS, a new architecture that reimagines how models learn and use memory. Unfortunately, it flew under everyone’s radar - but it shouldn’t have! So what's Atlas bringing to the table? ▪️ Active memory via Google’s so-called Omega rule. It https://t.co/A7XD4CXeS3
— TuringPost (@TheTuringPost) Jun 3, 2025
Windsurf CEO Varun Mohan says code generation is just the appetizer. Real engineering is debugging, testing, system design. As AI agents tap logs, databases, and browser context, “every aspect is going to be 10× more effective.” The real disruption is in engineering cognition. https://t.co/hi4nvn2SGQ
— vitrupo (@vitrupo) Jun 3, 2025
📢 New @a16z thesis: Market research is a $140B industry and most of that spend still goes to human surveys, panels, and agencies. For years, software was a rounding error. Now, AI is flipping the model on its head: https://t.co/gX3veWTFkx
— ZC25 (@zachcohen25) Jun 3, 2025
Dharmesh Shah is worth over a Billion Dollars. He grew Hubspot to a $30 BILLION company and invested in over 60 startups like CoinBase and Angel List. Best thing is He documents all his insights and puts them on the Internet. Here's his Guide To Growing Startups 🧵 https://t.co/Zmfybu4VKx
— Shawn Chauhan (@shawnchauhan1) Jun 2, 2025
The story of @windsurf_ai is wild. The company was founded in 2021. Windsurf was the third pivot of the company. They cannibalised $M's in revenue to pivot to the Windsurf product. Within months of launching Windsurf they hit $50M in ARR. And a reported $3BN acquisition by https://t.co/7fFxpyPBQP
— Harry Stebbings (@HarryStebbings) Jun 2, 2025
