Wide Research is our latest exploration in agent-agent collaboration. Built on our large-scale virtualization infrastructure, Manus can now autonomously dispatch a team of homogeneous Manus agents to work in parallel and aggregate the results. While building AI agents, we've https://t.co/wHrHa8AshT
— Yichao 'Peak' Ji (@peakji) Jul 31, 2025
People are wildly underestimating how entertaining it is for AI to write & produce a story FOR you. I put "Sam Altman tries to convince Kim K to promote GPT-5" into Showrunner (@fablesimulation) and got this 🤯 The product handles the script, images, voices, and animation. https://t.co/EGpzRyvkNL
— Justine Moore (@venturetwins) Jul 31, 2025
I have been warning about this for years. You have a short window now to download whatever field of human knowledge you want and be free to learn. Anna's Archive is one of the best sources for this You don't want your only access to knowledge neing dumbed down LLMs https://t.co/Y3PTwGP5RP
— Moriarty (@ThingsHiddenn) Jul 30, 2025
Microsoft just did a study on the 40 jobs most at risk by AI and the 40 most secure. A few really surprised, and many we own. https://t.co/ANO3E908Tk
— Codie Sanchez (@Codie_Sanchez) Jul 31, 2025
4 LLM fine-tuning libraries you need to know about as an AI engineer. 100% Opensource. https://t.co/CXsNy9FOE4
— Shubham Saboo (@Saboo_Shubham_) Jul 31, 2025
This study systematically reviews the evolution of AI and AI agent technologies, ---- Paper – arxiv. org/abs/2507.01376 Paper Title: "AI Agents and Agentic AI-Navigating a Plethora of Concepts for Future Manufacturing" https://t.co/khVWAB1CXw
— Rohan Paul (@rohanpaul_ai) Jul 31, 2025
To ultimately win, you have to be both first and last mover in your industry. - Peter Thiel https://t.co/6Y8CVLl0Ly
— Z Fellows (@zfellows) Jul 30, 2025
Context engineering > prompt engineering. Even @karpathy agrees Production AI needs memory, knowledge & RAG, Not just vibing with one-off prompts Stop coding blindly, start engineering context: https://t.co/dMaDOlD6Uk
— Mervin Praison (@MervinPraison) Jul 30, 2025
paul graham on good design https://t.co/gUTvHbzhWp
— keshav (@keshavchan) Jul 30, 2025
we are so back https://t.co/Oc9JIbdHbr
— tobi lutke (@tobi) Jul 29, 2025
Two years ago, this was stuff scene only in movies like Minority Report. https://t.co/7MvkqEKMuy https://t.co/0W5NzAQJcA
— sunny madra (@sundeep) Jul 29, 2025
The Matrix was a story about a simulation. The Matrix was about Aleph. https://t.co/deKP5Y6bwk
— Cristóbal Valenzuela (@c_valenzuelab) Jul 29, 2025
🇨🇳 China is ahead here too. China treats AI as normal study gear, not coursework cheating. Ministry plan and Beijing rules push AI literacy into every syllabus. Campuses run full‑spec DeepSeek on local servers, handing students free unlimited chat. Library card grants endless https://t.co/49v1B7tzQF
— Rohan Paul (@rohanpaul_ai) Jul 29, 2025
Tobi Lutke explains what the VCs who passed on Shopify got wrong Tobi recounts pitching Shopify to VCs on Sand Hill Road a few years after founding Shopify. Investors passed because they thought the addressable market was too small. At the time, there were about 40,000-50,000 https://t.co/qiPVzUdxym
— Startup Archive (@StartupArchive_) Jul 29, 2025
Most developers chunk first, embed second. But what if that's completely backwards? Let’s go through some advanced chunking techniques that can significantly improve your retrieval quality: 𝟭. 𝗦𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝗖𝗵𝘂𝗻𝗸𝗶𝗻𝗴 Instead of splitting at arbitrary character https://t.co/mpNT066Yom
— Femke Plantinga (@femke_plantinga) Jul 29, 2025
Here's BlockDL, a free & open-source GUI that lets you visually design Keras neural networks and learn ML. https://t.co/3r4qz02ghZ https://t.co/iTR2h4c8et
— François Chollet (@fchollet) Jul 29, 2025
Finally, a 45 page literature review of text embedding model, datasets, evaluation and training methods: https://t.co/EmVNNFR7KA
— Bo (@bo_wangbo) Jul 29, 2025
There are so many ways to improve Claude Code using hooks and sub agents. Here’s one example of many I’ve been testing. Ephemeral, scoped sub agents, Claude dynamically creates for a task and removes when the task is done. They have some advantages over generic Task sub agents https://t.co/MEuTl5duoE
— Eric Buess (@EricBuess) Jul 29, 2025
“Big opportunities in life have to be seized.” ~ Warren Buffett https://t.co/7Ezf7m8HAS
— Jon Erlichman (@JonErlichman) Jul 28, 2025
Definitely be reading this if you have not already. For me it's rare to recommend a bio (comp or person) while the story is still playing out and also from a reporter. This is a unique story and worth telling now and this is told well and respectfully. https://t.co/XeYDPx0hMg https://t.co/YGMsmVYCUu
— Steven Sinofsky (@stevesi) Jul 29, 2025
Less than 89 hours ago, Claude Code unlocked sub agents. Minds are blown. And people are already building with their own agentic AI dev team. 10 wild examples: https://t.co/OZ7yYijb1Y
— Min Choi (@minchoi) Jul 28, 2025
"No fluff. Pure gold mine. I am going to read both his books. "I am an AI founder. I found useful information that's immediately valuable." "He nailed every single mistake B2B SaaS is making and has a solution for each of them." "Finally, some practical advice on pricing AI https://t.co/XSU0e9wNEH
— Lenny Rachitsky (@lennysan) Jul 28, 2025
🚨 New investment announcement! We @a16z are leading Salient AI's $60M Series A. Salient builds AI for loan servicers. In the last 18 months, they've processed more than $1B in transactions and helped lenders cut handle times by 60%. (cc @arampell, @seema_amble) https://t.co/nXsgntEtIc
— Olivia Moore (@omooretweets) Jul 28, 2025
Streaming custom data alongside your LLM response is hard Data parts in @aisdk 5 provide an awesome (type-safe!) solution Docs now live https://t.co/dBS9leXT6W
— Nico Albanese (@nicoalbanese10) Jul 28, 2025
Never seen a CEO talking about AI so bluntly: “As a CEO myself, I can tell you, I’m extremely excited about it. I’ve laid off employees myself because of AI. AI doesn’t go on strike. It doesn’t ask for a pay raise. These things that you don’t have to deal with as a CEO.” I https://t.co/kX92FB7g3K
— Chubby♨️ (@kimmonismus) Jul 28, 2025
Heard @MadhavanSF speak at a @lightspeedvp event last year. Highest ROI per word I’ve encountered. Listen to this podcast, buy his book. https://t.co/x8X9LNqbFQ
— anand iyer (@ai) Jul 27, 2025
Updated our RAG agent recipe to v5 https://t.co/tvbXBLZ6xj
— Nico Albanese (@nicoalbanese10) Jul 27, 2025
Lovable is going FULL STACK. Soon, you'll be able to add APIs, databases, or even Claude/Stripe directly into your build. No more work arounds. Just drop it in and it works. Imagine the possibilities: • One-click OpenAI API? • Custom backends? • MCPs baked in? The future https://t.co/IdZ1PD6fX0
— damien (@damienghader) Jul 27, 2025
I have been fine-tuning LLMs for more that 2 years now! Here are the top 5 LLM fine-tuning techniques, explained with visuals:
— Akshay 🚀 (@akshay_pachaar) Jul 27, 2025
Interview with Tesla Owners SV https://t.co/uecwAwGSAx
— Elon Musk (@elonmusk) Jul 27, 2025
The Chinese have a $5,900 robot that can do backflips. This will continue to get better in capability. I don't know how the US can compete in this arena. It just seems like the race is pretty much over before it has even started. https://t.co/GxpIJ7kaeu
— Carlos E. Perez (@IntuitMachine) Jul 26, 2025
Here is another example of how I plan and code with Cursor. - Make a plan 1. Link to trace function calls or architecture spec. 2. Add relevant docs to the prompt 3. Reference code examples 4. Save the plan to a .md file 5. Discuss your changes (you don't have to answer all the https://t.co/sqf9MHB25c
— Kevin Kern (@kregenrek) Jul 26, 2025
Many of the big labs are working in similar directions, including Google. This looks powerful. https://t.co/KuFxTZZoyl https://t.co/ZtMHaPxngB
— Andrew Curran (@AndrewCurran_) Jul 27, 2025
MCP is getting so good! https://t.co/DzW7djTf35
— Santiago (@svpino) Jul 26, 2025
Claude Code sub-agents is probably the easiest way to build complex, custom agentic systems. Not just for code, but for anything you can imagine. Watch me build a multi-agent deep research system with Claude Code subagents. Chain subagents with /commands for reliability. https://t.co/8zGNPMSSfV
— elvis (@omarsar0) Jul 26, 2025
MetaMCP turns multiple MCP servers into one unified endpoint. Organize servers into namespaces, filter tools, and add middleware layers with Docker deployment support. 100% opensource. https://t.co/Rd0evTVz03
— Shubham Saboo (@Saboo_Shubham_) Jul 25, 2025
Promise fulfilled! Took me 9 days of craziness but my Context Engineering with DSPy tutorial is out now on YouTube. My next goal: getting my sleep schedule back on track. https://t.co/xbfJlJ3iw8 https://t.co/RLAUbuu0Ub
— AVB (@neural_avb) Jul 25, 2025
I built 7 custom subagents for Claude Code to help me ship faster! 1. code-refactorer 2. prd-writer 3. project-task-planner 4. vibe-coding-coach 5. security-auditor 6. frontend-designer 7. content-writer (Try them yourself from the link below) https://t.co/ohcdF78I1e
— Ian Nuttall (@iannuttall) Jul 25, 2025
The headline here is that sub agents are a great abstraction boundary for context engineering. If you want to learn more about this exciting new feature, check out our docs https://t.co/Ckt8N14kzX https://t.co/Mos7kJjAIB
— Adam Wolff (@dmwlff) Jul 25, 2025
This is an awesome article. The best part is their note to “build around the KV cache”. If your system prompt remains consistent, your tools remain constant, and you always append to conversation json… you will hit the KV cache often. Cutting down cost and latency. https://t.co/KpBtJvsdSX
— AVB (@neural_avb) Jul 25, 2025
AI doesn't search like humans do. It wants paragraphs, not keywords. It needs 1000 results, not 10. This is why traditional search engines fail AI systems. They were built for lazy humans who can't type full sentences. The future of search isn't about clicks, it's about https://t.co/tZBep1Bt7b
— jason liu (@jxnlco) Jul 25, 2025
Microsoft just released their list of the 40 jobs most AI-applicable and the 40 jobs least AI-applicable. This list may have overlap with replaceability and job transformation. (Phlebotomists, you've got nothing to worry about.) https://t.co/WHHMIqFb21
— Allie K. Miller (@alliekmiller) Jul 23, 2025
Congrats to Lovable! The world's fastest-growing businesses are built on Stripe. https://t.co/t0lEi2vDDd
— Patrick Collison (@patrickc) Jul 23, 2025
An agent who can do work for you is a big big idea https://t.co/WTtStTbc7W
— Garry Tan (@garrytan) Jul 23, 2025
Today, we're excited to introduce you to Kafka, the first AI employee. Kafka is the first step towards our goal of creating an AI worker that is virtually indistinguishable from a remote human employee. Kafka is the new state-of-the-art on the GAIA Level 3 benchmark which https://t.co/oCHmCigJnd
— Brainbase Labs (@BrainbaseHQ) Jul 23, 2025
Craft is truly caring about what you make. Software or sushi, either way it is about absolute dedication to creating something people want. The Lesson of Craft https://t.co/XajFJYWUei https://t.co/5aOOR4y8YW
— Garry Tan (@garrytan) Jul 23, 2025
everyone in AI should start learning JSON prompting It might look complicated at first, but once you have a closer look at the structure, it’s actually much more clearer everything is organized, camera, lighting, shot details, transitions, motion, each in its own section, so https://t.co/1WundMCQuI
— el.cine (@EHuanglu) Jul 23, 2025
Here's my conversation with @demishassabis, CEO of Google DeepMind, all about the future of AI & AGI, simulating biology & physics, video games, programming, video generation, world models, Gemini 3, scaling laws, compute, P vs NP, complexity, energy (solar & fusion), and much https://t.co/VhbqmBQHje
— Lex Fridman (@lexfridman) Jul 23, 2025
I've been working on a new type of advanced agent for months now (will be sharing more info soon). I tried integrating pretty much every MCP service on the market. They all sucked, except for @composiohq. Very cool that they've now raised $25M. This company is going to go far! https://t.co/ERDMBVGXI7
— Matt Shumer (@mattshumer_) Jul 22, 2025
Former $AMZN employee breaks down the cloud market's competitive shifts ( $MSFT, $GOOGL, $NVDA, $CRWV ): - The expert believes $AMZN made a strategic misstep by not aligning with $NVDA early in the generative AI wave, which left it without a competitive GPU offering as the https://t.co/bt1TH2S3Rf
— AlphaSense (@AlphaSenseInc) Jul 22, 2025
Congrats to our friends from @composiohq! https://t.co/rKCCpwyknW
— Jerry Liu (@jerryjliu0) Jul 22, 2025
In this report Anthropic argues that, in order to stay at the frontier of AI development, the United States must be prepared to operate at least 50GW of power capacity for AI workloads by 2028. https://t.co/b7FDWXy5fS https://t.co/Da7UPJF9lN
— Andrew Curran (@AndrewCurran_) Jul 22, 2025
Nvidia is killing it these days and sharing in the open for everyone's benefit! Currently second big organization on HF by number of repos created in the past 12 months (average a new repo every day) and crossed 30,000 followers (second in the US just after Meta). https://t.co/z0mlzVeWhS https://t.co/YnBMfNyEFb
— clem 🤗 (@ClementDelangue) Jul 21, 2025
VEO3 → Corona party in a bottle. Utilizing JSON prompting. + ChatGPT: Prompt Refinement + Veo3: Video Gen (FULL PROMPT in thread) #veo3 #promptshare https://t.co/tE9HJ9MFdc
— Rory Flynn (@Ror_Fly) Jul 21, 2025
Context window size is an important limitation for all LLMs. This limitation becomes particularly apparent in large engineering organizations using models for multi-turn agent interactions. The common approaches to get around this are conversation history cut-offs or https://t.co/z7McCsTGxf
— Factory (@FactoryAI) Jul 21, 2025
These agentic teams based models like Grok Heavy, the Gemini Deep Think that just won gold, and the next gen from OpenAI are all going to use about fifteen times more tokens than current systems. This is why Pro plans are north of $200. Essentially; Jensen wins again. https://t.co/aI4jOSCeoM
— Andrew Curran (@AndrewCurran_) Jul 21, 2025
Peter and Dave have an extraordinary show on the future of everything and abundance ! I was honored to be on this one https://t.co/CnAj0DTuJc
— Eric Schmidt (@ericschmidt) Jul 21, 2025
Amazing article by @rasbt comparing 6 LLMs Architecture, this Kimi 2 @Kimi_Moonshot is phenomenal, 1 TRILLION → 54B active! MoE is wild - more knowledge, same compute. I made a Claude artifact, you can play with it to compare models, plus a glossary of terms used in this https://t.co/xa9z17lbpV
— Unclecode (Hossein) (@unclecode) Jul 20, 2025
>OpenAI was founded in 2015. The seminal transformers paper “Attention is All You Need” wasn’t published until 2017. You have to be in the water. https://t.co/ZISuilCoGv
— Alexey Guzey (@alexeyguzey) Jul 21, 2025
Such a great conversation with @AravSrinivas, CEO of Perplexity. AI Browsers, agents, white collar future, and competition. Watch it if you haven't already. https://t.co/SEaUz3tsOQ
— Matthew Berman (@MatthewBerman) Jul 20, 2025
OpenAI dropped a massive resource on getting started with deep RL(RL + deep learning) link below https://t.co/2k2qx4RoHQ
— verrsane (@verrsane) Jul 20, 2025
A beautiful visual blog, where you can change values, interact, and see what each head does exactly inside the transformer. https://t.co/gxvJlXRoNK
— Pramod Goyal (@goyal__pramod) Jul 20, 2025
🚨China, USA, Japan aren’t just building robots. They’re training them to fight, film, cook and even making cyber horses. Here are 10 wild examples you won’t believe are real: 1. Kawasaki built a 4-legged robot, feels straight out of sci-fi. https://t.co/WeacJIjzUi
— The AI Colony (@TheAIColony) Jul 19, 2025
Eric Schmidt says traditional user interfaces are going to go away. The WIMP model (windows, icons, menus, pull-downs) was built 50 years ago. In the age of agents, UI becomes ephemeral. Generated on demand, shaped by intent, not layout. https://t.co/CFS7moMoYz
— vitrupo (@vitrupo) Jul 20, 2025
Still blown away by Flux Kontext. They just need to improve the quality a bit. Besides that it’s a fantastic model. Being able to create different consistent shots based on a single image is amazing. https://t.co/WUQVkhiMuI
— Halim Alrasihi (@HalimAlrasihi) Jul 18, 2025
This is basically Cursor over your research papers 🧑💻 Besides asking research questions, help read (and write) commands and code to do deeper analysis on implementation details. Check it out below 👇 https://t.co/CakNt0M7yP
— Jerry Liu (@jerryjliu0) Jul 18, 2025
From GPT to MoE: I reviewed & compared the main LLMs of 2025 in terms of their architectural design from DeepSeek-V3 to Kimi 2. Multi-head Latent Attention, sliding window attention, new Post- & Pre-Norm placements, NoPE, shared-expert MoEs, and more... https://t.co/oEt8XzNxik
— Sebastian Raschka (@rasbt) Jul 19, 2025
Build a deep research agent over offline and online data with AWS Bedrock 🤖🪨 Check out this @aws guide on how to build an agent researcher that can pull from online data (Github, TechCruch, Duckduckgo search) and offline data, all within the AWS ecosystem! P.S. As an aside - https://t.co/TqSiNUoEjA https://t.co/66U6VHEvLy
— Jerry Liu (@jerryjliu0) Jul 18, 2025
We have to come to the realization that LLMs think in ways that are different from humans. In fact, their cognition may be complementary to ours (just as computers already are). https://t.co/FMtluw6ox3
— Carlos E. Perez (@IntuitMachine) Jul 19, 2025
“There's an unspoken covenant that as a founder, you go down with the ship. For better or worse, it's changed a bit over the last year and I think it's disappointing, to be honest.” Enough said. This show is everything and more on: - What really happened behind the scenes - https://t.co/qaY7MVwgIy
— Harry Stebbings (@HarryStebbings) Jul 18, 2025
2025.29: What It Takes to Change the Web The best Stratechery content from the week of July 14, 2025, including the economic future of the web, Greatest of All Talk in Las Vegas, and how we cool computers. https://t.co/Tm0gJUuFaU
— Stratechery (@stratechery) Jul 18, 2025
Never use Pure White (FFFFFF). It can lead to eye strain and looks really harsh. Instead, use these alternatives: https://t.co/UXTXcEjnpV
— Ruben (@rubenhume) Jul 17, 2025
Apple has published a 27-page long paper called "Apple Intelligence Foundation Language Models - Tech Report 2025" https://t.co/fNlERIc22q
— Aaron (@aaronp613) Jul 17, 2025
Aaron Levie: How AI Will Transform Work "In ten years from now when we all have AI agents running around, we will just look back to today and be like, how did we function? Like you spent two weeks to decide the message for the marketing campaign?" — @levie https://t.co/9HR0oh5n9N
— AI Insights (@plzaccelerate) Jul 17, 2025
Agents to extract structured tables from any PDF in seconds ⚡️ A lot of PDFs (financial reports, research papers, invoices), have complex charts and tables that require a lot of human effort to format into a tabular representation (e.g. in Excel, CSV, or Pandas). We added a https://t.co/HE7quJHa3V https://t.co/QX6NmCZG2R
— Jerry Liu (@jerryjliu0) Jul 17, 2025
Surprise! Operator and deep research had a baby, and it’s called ChatGPT Agent 🚀🚀 It's a research agent with full access to a computer that can do a host of things that neither of its parents could do on their own: Navigate complex multi-step workflows: Read all of your https://t.co/o2fxaly1bd
— Dan Shipper 📧 (@danshipper) Jul 17, 2025
OPENAI UNVEILS AGENT THAT CAN MAKE SPREADSHEETS AND POWERPOINTS - WSJ https://t.co/yRzGlYCFmI
— Evan (@StockMKTNewz) Jul 17, 2025
we missed a banger paper in the grok4/k2 drop noise guys. these guys > look for optimal ways to select data mixes to get max improvement on a model given a target domain. > do multimodal validation > show good extrapolation accuracy (testing on 1.4B and predicting on 8B) https://t.co/6EPjxlRAZJ
— tokenbender (@tokenbender) Jul 17, 2025
The 3 phases of the Foundation Sprint (aka, the missing manual for early-stage founders): 1. The basics—who's your customer, what problems are you solving for them, who's the competition, what's your edge 2. Differentiation—how will you specifically differentiate from the https://t.co/agD8qiAj4f https://t.co/L49hXRLckH
— Lenny Rachitsky (@lennysan) Jul 16, 2025
I think I am cooking the Context Engineering with DSPy tutorial hard. The blueprint. 👇🏽 https://t.co/PnCPOIawuU https://t.co/lYft7w65FV
— AVB (@neural_avb) Jul 16, 2025
i had an idea for a small self-contained feature inside of @CoraComputer this morning 2 hours later i just pushed a PR for @kieranklaassen. 100% done by Claude Code, i have never looked at the codebase he's spent a ton of time architecting Cora's repo to make it super simple
— Dan Shipper 📧 (@danshipper) Jul 16, 2025
Solo founders + agents are here https://t.co/aR1ORlcAD6
— Ben Lang (@benln) Jul 16, 2025
Layoffs at Scale one month after Meta investment, per @shiringhaffary & @rachelmetz https://t.co/Wx8oPaOEg7
— Kate Clark (@KateClarkTweets) Jul 16, 2025
While most organizations are just starting to deploy gen AI at some scale, AWS wants to be the best place for Agent development. https://t.co/Qyye8MJhLP
— Ben Bajarin (@BenBajarin) Jul 16, 2025
Recommended listening for everyone. Research Scientist from OpenAI. Give your 30 mins to this. Great way to look at things, and how can you create wealth in new economy. https://t.co/IlvJKHb1xy
— Gabbar (@GabbbarSingh) Jul 16, 2025
Fusion of 𝗥𝗔𝗚 (Retrieval Augmented Generation) and 𝗖𝗔𝗚 (Cache Augmented Generation). How can you benefit from it as AI Engineer? Few months ago there was a lot of hype around a technique called CAG. While it is powerful to its own extent, the real magic happens when you https://t.co/x9nNVpZnTh
— Aurimas Griciūnas (@Aurimas_Gr) Jul 16, 2025
Announcing a new Coursera course: Retrieval Augmented Generation (RAG) You'll learn to build high performance, production-ready RAG systems in this hands-on, in-depth course created by https://t.co/R0m408f8CA and taught by @ZainHasan6, experienced AI and ML engineer, researcher, https://t.co/KZjnENFl81
— Andrew Ng (@AndrewYNg) Jul 16, 2025
At today’s @awscloud Summit NYC, we unveiled a new Bedrock capability called AgentCore – a set of powerful building blocks that will change how you can deploy agents into production in a secure, scalable, and flexible way. https://t.co/WVzY30ieoP https://t.co/nRmizWEmnB
— Andy Jassy (@ajassy) Jul 16, 2025
S3 is a vector DB now FYI: https://t.co/ucdl1ClLz1
— Randall Hunt (@jrhunt) Jul 16, 2025
Chain of Thought (CoT) monitoring could be a powerful tool for overseeing future AI systems—especially as they become more agentic. That’s why we’re backing a new research paper from a cross-institutional team of researchers pushing this work forward. https://t.co/dmM8azLcWE
— OpenAI (@OpenAI) Jul 15, 2025
Modern reasoning models think in plain English. Monitoring their thoughts could be a powerful, yet fragile, tool for overseeing future AI systems. I and researchers across many organizations think we should work to evaluate, preserve, and even improve CoT monitorability. https://t.co/MZAehi2gkn
— Bowen Baker (@bobabowen) Jul 15, 2025
A restaurant in Tokyo where wagyu meats are grilled infront of you https://t.co/y2muz0x6Yt
— Science girl (@gunsnrosesgirl3) Jul 15, 2025
The Bitter Lessons Behind Kimi Researcher's taste. Ideas are cheap, get things done is what truly matters. https://t.co/84nydP3lbN
— Crystal (@crystalsssup) Jul 15, 2025
insanely high signal blog post by fal research team. https://t.co/6xyXhSK1XZ
— maharshi (@mrsiipa) Jul 15, 2025
a guy created a dataset of 50 books from London 1800-1850 for LLM training. no modern bias. it’s actually super cool to see what can be trained on it! https://t.co/iAM5enRqeL
— ℏεsam (@Hesamation) Jul 14, 2025
"But another way to describe Salesforce is as a playbook...and Salesforce, via the product’s vocabulary, features, and defaults, is a loose encoding of those recommended processes." Great post by @bennstancil on the fact that SaaS is a wrapper around a static ontology. https://t.co/IVpm5tdDjT
— Peter Wilczynski (@petewilz) Jul 14, 2025
I wrote this thread a while ago detailing how I write my initial prompts for a game project. I think it’s still relevant now, especially with more powerful models. Hope it’s helpful if you want to “vibe code” some games. Have fun and enjoy the process! https://t.co/CuIjLtffm7
— Danny Limanseta (@DannyLimanseta) Jul 14, 2025
🚨NVIDIA CEO just redpilled the MIT study claiming AI makes people dumber > “I haven’t read the MIT research lol” > “I literally use AI every day and my cognitive skills are actually advancing” > most of his job as CEO = asking good questions > “I’m not asking a machine to teach https://t.co/u1dFZew5ep
— NIK (@ns123abc) Jul 14, 2025
this is wild… Grok 4 is now basically a Wall Street intern that never sleeps. • Reads filings • Analyzes earnings calls • Flags red/green signals • Suggests sector rotations prompt + live example inside 👇
— Alex Hughes (@alxnderhughes) Jul 13, 2025
Btw if you're learning how to build LLMs from the ground up, there's now a 17h companion video course for my LLMs From Scratch book on Manning: https://t.co/GwwQBKOPAk It follows the book chapter by chapter, so it works great either as a standalone or code-along resource. It's https://t.co/AdZjyeqLsy
— Sebastian Raschka (@rasbt) Jul 13, 2025
ICONIQ published its 67 page State of AI report for 2025 They interviewed 300+ executives at AI companies to understanding what has changed & what hasn't in the world of AI Hint: AI is eating labor, but this isn't a smooth ride Sharing 5 notable takeaways ⤵️
— Rahul Mathur (@Rahul_J_Mathur) Jul 12, 2025
Co-founder and CEO of Stripe's advice to ambitious young people: https://t.co/hoeIvO4Oew
— Jay Yang (@Jayyanginspires) Jul 12, 2025
12 MONOPOLIES SHAPING THE FUTURE 1. $NVDA powers AI 2. $META rules social 3. $GOOGL owns search 4. $PLTR became AI’s OS 5. $NFLX defined streaming 6. $AXON runs public safety 7. $TSLA drives real-world AI 8. $ASML controls chipmaking 9. $TSM owns chip production 10. $AMZN built
— Shay Boloor (@StockSavvyShay) Jul 12, 2025
Here’s what I think will happen: 1. Successful companies will have fewer employees 2. Everyone (PM, Eng, etc) will be expected to build and to have product sense 3. The PMs who thrive in this new reality will have a very different skill set from the typical PM today 4. The https://t.co/sTvSVIcTBH
— Peter Yang (@petergyang) Jul 12, 2025
Adaptive Form Extraction 🤖📑 Form understanding (e.g. tax forms, patient intake, certifications, questionnaires) is a huge use case for AI agents, but defining a good output schema is painful: 1) it can be very tedious/custom, and 2) each form has its own values. 💡 Instead of https://t.co/mNCiIenVR5
— Jerry Liu (@jerryjliu0) Jul 12, 2025
Introducing GenAI Processors ✨ An open source library to build real-time projects easily, with cool features such as stream-based I/O and chaining, modularity, composability, and more GitHub: https://t.co/FGxKcn8hNB Blog: https://t.co/vszE6mFm4P https://t.co/OgFliV1Q7k
— Omar Sanseviero (@osanseviero) Jul 10, 2025
Always enjoy talking to @dflieb on and off camera about everything: search, AI, product. In this video, we discuss why the next big bet is to nail agents and the browser. Thanks to @garrytan for inviting me. https://t.co/6aUxomjF5R
— Aravind Srinivas (@AravSrinivas) Jul 11, 2025
Here @elonmusk proves once again that his methods work. Assemble a small team of world-class engineers who work at a hardcore pace. An excerpt from the Elon Book (coming soon!) https://t.co/qVBoFkQPu5 https://t.co/Ib33ucmhCP
— Eric Jorgenson 📚 ☀️ (@EricJorgenson) Jul 11, 2025
The future business model of software will likely be seats for the user and consumption for the AI. But seats have always been capped by the number of workers you can sell to, whereas AI Agent consumption is unbounded. So this will be the bigger component over time. https://t.co/itpEKdIG4O
— Aaron Levie (@levie) Jul 12, 2025
Very excited to share that @windsurf_ai co-founders @_mohansolo & Douglas Chen, and some of their talented team have joined @GoogleDeepMind to help advance our work in agentic coding in Gemini. Welcome to our new team mates from Windsurf! https://t.co/obSwzXy6t5
— koray kavukcuoglu (@koraykv) Jul 11, 2025
NEWS: AMAZON IS LAUNCHING AI AGENT MARKETPLACE WITH ANTHROPIC > marketplace model will allow startups to charge customers for AI agents > startups to directly offer AI agents to AWS customers > marketplace will help aws and anthropic reach more customers > and attract developers https://t.co/Ll3BHYz8JM
— NIK (@ns123abc) Jul 11, 2025
In a survey by Oppenheimer & Co. related to user satisfaction with Google's AI Mode, of people who have used AI Mode and pay for ChatGPT: - 82% find AI Mode results more helpful than Google Search - 75% find AI Mode results more helpful than ChatGPT This echoes what I wrote https://t.co/4jsxWxcStq
— Eric Seufert (@eric_seufert) Jul 11, 2025
Agentic Document Extraction now supports field extraction! Many doc extraction use cases extract specific fields from forms and other structured documents. You can now input a picture or PDF of an invoice, request the vendor name, item list, and prices, and get back the extracted https://t.co/dxguIt97OT
— Andrew Ng (@AndrewYNg) Jul 10, 2025
My talk at YC Startup School on how to build AI startups. I share tips from @AI_Fund on how to use AI to build fast. Let me know what you think! https://t.co/6AmVXgP4lc
— Andrew Ng (@AndrewYNg) Jul 10, 2025
How bootstrapped solo founder @MS_BASE44 has been able to compete head-on with heavily funded competitors like Lovable, Bolt, Replit, Vercel, etc. "It's a different ball game because even if you're a very small team, and even if you're solo, you're literally managing teams of AI https://t.co/gmJprWEPf0 https://t.co/6qyGuxgKdI
— Lenny Rachitsky (@lennysan) Jul 10, 2025
xAI gave us early access to Grok 4 - and the results are in. Grok 4 is now the leading AI model. We have run our full suite of benchmarks and Grok 4 achieves an Artificial Analysis Intelligence Index of 73, ahead of OpenAI o3 at 70, Google Gemini 2.5 Pro at 70, Anthropic Claude https://t.co/Vc9781SIzd
— Artificial Analysis (@ArtificialAnlys) Jul 10, 2025
This is all literally all you need to train a judge https://t.co/9goYczRYeM
— Casper Hansen (@casper_hansen_) Jul 9, 2025
GROK 4 HEAVY > scale-up test-time compute by 10x > spawns multiple agents in parallel > all agents work independently then compare work “like a study group” > figure the best result > SOLVED https://t.co/B8LIL1F4y5
— NIK (@ns123abc) Jul 10, 2025
7-9-2025 ($) •Apple + Anthropic? •Apple's Fall •Apple's Options https://t.co/xqoh9YWgTz
— Stratechery (@stratechery) Jul 9, 2025
"Bro trust me, AI-generated meshes will never be production-ready" The AI-generate mesh: https://t.co/2Xtm6Je6nZ
— Emm | scenario.com (@emmanuel_2m) Jul 8, 2025
OpenAI CEO shared how to build a money-making startup: https://t.co/svQAq2Ywqn
— Meera | AI Tools & News (@Meer_AIIT) Jul 8, 2025
I got permission to publish this massive EI Evals FAQ (PDF). It's like a bible for AI engineers and AI PMs. @HamelHusain and @sh_reya answer the most common questions they got while teaching 700+ students. And share 30+ free videos, posts, and resources: 🧵 https://t.co/rJrVi3Pr8x
— Paweł Huryn (@PawelHuryn) Jul 8, 2025
This is what the ideal grocery store looks like. Minimally processed (NOVA Group 1) food only (no "edible food-like substances"), organic, local, fresh. Food should not be more complex than this, yet I don't believe this exists. https://t.co/joHUslcdM3
— Andrej Karpathy (@karpathy) Jul 8, 2025
What you did last year is not good enough to keep your job this year. You need to Re qualify every year. https://t.co/bUsPlqxdgt
— Shane Parrish (@ShaneAParrish) Jul 8, 2025
Rarely do you see graphs that so clearly indicate a paradigm shift https://t.co/uGKAt8US6P
— Prakash (Ate-a-Pi) (@8teAPi) Jul 8, 2025
Three scientists (Christian Catalini, Jane Wu, Kevin Zhang) have written a very interesting article in the Harvard Business Review whose views I share; a scientist wrote to me via email and talked to me about the article. The message: the potential danger for the labor market is https://t.co/DjKaKTF47Q
— Chubby♨️ (@kimmonismus) Jul 8, 2025
I think of this Peter Thiel clip often. Thiel discusses how he sold $META at IPO; underestimating just how easy it would be for the company to scale from a $100B market cap to $1T. While this conversation is specifically about big tech, I believe it can apply to any company https://t.co/9GKpUBBKJY
— Jack Prescott (@JackPrescottX) Jul 8, 2025
Today we published a targeted transparency framework for frontier AI development. Our framework focuses on major frontier model developers while exempting startups and smaller developers to avoid burdening the broader ecosystem. https://t.co/UrU8Vjtsm4
— Anthropic (@AnthropicAI) Jul 7, 2025
I wrote a blog post comparing all the recent video models on Replicate 😅 https://t.co/kKuEnoVNEX https://t.co/JBcV3GBfhg
— fofr (@fofrAI) Jul 8, 2025
It took me a little while but I am finally reposting this tutorial on my (new) personal website / blog. It should be easier to share and read then a X article, let me know what you think! maximerivest[.]com https://t.co/mf5SIXDyUn
— Maxime Rivest 🧙♂️🦙🐧 (@MaximeRivest) Jul 7, 2025
The System Prompts on Meta AI's agent on WhatsApp have been leaked. It's a goldmine for human manipulative methods. Let's break it down. Comprehensive Spiral Dynamics Analysis of Meta AI Manipulation System BEIGE Level: Survival-Focused Manipulation At the BEIGE level, https://t.co/hpHXwMlWcL
— Carlos E. Perez (@IntuitMachine) Jul 5, 2025
Kevin O'Leary shares the most important lesson he learned from Steve Jobs‼️‼️ https://t.co/VoNEonsn35
— Leaders 𝕏 Junction (@LeadersJunction) Jul 6, 2025
High pricing is a new moat https://t.co/6TpjTo98eT
— Garry Tan (@garrytan) Jul 5, 2025
People don't understand how far you can take this stupidly simple workflow. Trigger → AI Agent → Output. That's it. Three fucking nodes. But this basic structure can literally become anything your degenerate heart desires: Email arrives → AI analyses and responds → https://t.co/sX0wCzfUyf
— Nozz (@NoahEpstein_) Jul 5, 2025
This is the most powerful Deep Research AI prompt I've ever used It can literally make you thousands of dollars The AI will research your niche and give you a DETAILED plan on how to build $$ making apps (no coding required) Drop what you're doing and plug this into AI now:
— Alex Finn (@AlexFinnX) Jul 5, 2025
Small Language Models (SLMs) are the future of Agentic AI, claim @NVIDIA researchers. Moreover, they offer a method for converting existing agent systems from using LLMs to SLMs that could work in practice. Here are the details: https://t.co/1g7LKYfKjI
— TuringPost (@TheTuringPost) Jul 5, 2025
Wrote up a few thoughts on Cursor's new $200/month Ultra plan and changes to their $20/month Pro plan https://t.co/2gAjMXN6CX https://t.co/78UsKIrYI4
— Simon Willison (@simonw) Jul 5, 2025
Is this AI agent startup the NEW n8n killer? How to set up vibe marketing automations with simple prompts (not crazy workflows) in 40 minutes https://t.co/1IRny8vd8O
— GREG ISENBERG (@gregisenberg) Jul 2, 2025
Google dropped a 68-page guide on prompt engineering.. packed with tips from beginner to advanced → kaggle .com/whitepaper-prompt-engineering While you're at it, try this AI agent to automate your tasks: https://t.co/qCCzlStsES https://t.co/2u5EczinCo https://t.co/RGLamEjXxT
— Nelly; (@nrqa__) Jul 4, 2025
Stop throwing better models at your RAG problems. Start optimizing what happens BEFORE retrieval instead. The secret isn't just in better embeddings or fancier models - it's in how you optimize queries before they hit your database. Here are 4 pre-retrieval techniques to https://t.co/tymFfj65VW
— Victoria Slocum (@victorialslocum) Jul 3, 2025
This GitHub repo is a gold mine for EVERY data scientist! DS Interactive Python repo has interactive dashboards to learn statistics, ML models, and other DS concepts. Topics include PCA, bagging & boosting, clustering, neural networks, etc. Fully open-source and free! https://t.co/T8Ll2SXmWo
— Daily Dose of Data Science (@DailyDoseOfDS_) Jul 3, 2025
What a gem. Eliad Shahar curated an insane @n8n_io database with 2,000+ AI workflows (and growing fast) across 365 APIs & services! 🤯 Also comes with blazing-fast docs for instant search. repo in 🧵↓ https://t.co/g1DDX7cYRw
— Charly Wargnier (@DataChaz) Jul 3, 2025
Marc Andreessen on Mark Zuckerberg’s founder “superpower” “There are certain fields in which you really get to see somebody’s core personality. You get to see their core attributes, their core virtues, their core vices, and their core weaknesses. And it’s really only under https://t.co/xTQUYMReMI
— Startup Archive (@StartupArchive_) Jul 2, 2025
An early look at the future of AI Studio (home page) https://t.co/1Wwciq34nF
— Logan Kilpatrick (@OfficialLoganK) Jul 1, 2025
From @zebriez essay “Flounder Mode” https://t.co/hVBpMXPmy7
— Patrick OShaughnessy (@patrick_oshag) Jul 2, 2025
This is the Cursor moment for design. It’s called MagicPath and it’s an infinite canvas where you literally chat your way to production-ready designs. Figma completely redesigned. 5 wild examples + how to try free 👇: https://t.co/tnIoB3yoiE
— Alvaro Cintas (@dr_cintas) Jul 1, 2025
BREAKING PROMPTING NEWS! Now there's a new development skill that needs to be developed! Working with Claude Code hooks! This is a new kind of control that can be leveraged in unexpected ways! How does this change the way we prompt? Let me explain why it's a radical change! https://t.co/xHViojFXll
— Carlos E. Perez (@IntuitMachine) Jul 1, 2025
Error analysis is all you need. These guys literally teach us how to do it, and wrote an entire chapter on it! (35% discount: https://t.co/3botO05aw6) Analyze: This stage covers identifying erros and understanding why they occur. By collecting samples and categorizing failure https://t.co/fvGNZGBkPB
— Eugene Yan (@eugeneyan) Jul 2, 2025
