Google just made a $2,000 AI education free (and most people will still choose to stay illiterate). They quietly launched Google Skills, a hub with 3,000 technical modules that replaces "prompting" fluff with actual DeepMind research workflows. You can now access the exact https://t.co/n38Ag0b1ea
— Aadit Sheth (@aaditsh) Dec 30, 2025
Integrating 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗥𝗔𝗚 Systems via 𝗠𝗖𝗣 👇 If you are building RAG systems and packing many data sources for retrieval, most likely there is some agency present at least at the data source selection for retrieval stage. This is how MCP enriches the evolution of https://t.co/l8PVIc8Jn4
— Aurimas Griciūnas (@Aurimas_Gr) Dec 31, 2025
Sandboxes, but also sharing that sandbox context and memory between agents. https://t.co/uDUf7Qxxnv
— Ivan Burazin (@ivanburazin) Dec 31, 2025
I'm very excited about the possibilities of AI for improving learning and education. Gemini shows some of this potential, and it's only going to get better. https://t.co/VTZbcTg8P8
— Jeff Dean (@JeffDean) Dec 31, 2025
1/ Since @JayaGup10 and I wrote about context graphs, the response has been huge...but I've also noticed a few misconceptions worth addressing. The tldr: context graphs aren't a graph database or structured memory. They require a fundamentally different approach to schema and
— Animesh Koratana (@akoratana) Dec 31, 2025
Spent the last 2 hours going through this and one thing is clear: If you spend 7 days mastering this repository, you can land a 6 figure job in 2026. Throw the GitHub link into Gemini, have it teach you step by step at a high level (to understand terminology and how to use the https://t.co/W47E8amwOt
— vas (@vasuman) Dec 31, 2025
probably the single most helpful vibe coding tip i can offer that i havent seen anyone talk about yet: in development, log ~every execution step out with the idea of helping your LLM debug its own code help your LLM help you. one upfront intentional investment in logging, then https://t.co/yMMarHiijt
— swyx 🇸🇬 (@swyx) Dec 30, 2025
🚨NEW Whitepaper on AI Agents by OpenAI The maker of ChatGPT shares how it builds AI Agents. Get the 34-page white paper here: https://t.co/uBPS87B1dN
— 🔥 Matt Dancho (Business Science) 🔥 (@mdancho84) Dec 30, 2025
Great explanation from a Reddit user on why Claude Code is so appealing for non-coding knowledge work. https://t.co/ozn0u4L1XV
— Peter Yang (@petergyang) Dec 29, 2025
🚨 BREAKING: Microsoft launches a free Python library that converts ANY document to Markdown Introducing Markitdown. Let me explain. 🧵 https://t.co/OqIuHg9sy7
— 🔥 Matt Dancho (Business Science) 🔥 (@mdancho84) Dec 29, 2025
Holy shit... Google DeepMind just exposed why everyone's been doing AI reasoning wrong. The AlphaGo team doesn't use chain-of-thought. They use parallel verification loops. And it's destroying every "advanced reasoning" technique you've heard about. Here's what they discovered https://t.co/BrL3SvtB7z
— Jainam Parmar (@aiwithjainam) Dec 29, 2025
Bot made $900 -> $208k in 3 months on polymarket one of the most talked about bots on polymarket right now farmed $208,521 pure profit in a quarter focus exclusively on esports mainly LoL and Dota 2 bot monitors live game parses stream and official data faster than market https://t.co/YTueu8U3Dm
— Archive (@ArchiveExplorer) Dec 29, 2025
"The purest form of ambition and most intense source of energy in the business is the founder CEO. Every next concentric circle of management beyond the founder CEO has the potential to be an order of magnitude drop off in intensity. And that is f*cking dangerous. You don't https://t.co/46JjJc5sJQ https://t.co/ok1Dj8hcNC
— Lenny Rachitsky (@lennysan) Dec 28, 2025
BREAKING 🚨: Manus AI got acquired by Meta. “We’re excited about what the future holds with Meta and Manus working together and we will continue to iterate the product and serve users that have defined Manus from the beginning.” https://t.co/Mae6kdDq0t https://t.co/LQXVbLSTuT
— TestingCatalog News 🗞 (@testingcatalog) Dec 29, 2025
Andrej Karpathy posted about the new programmable layer developers need to master: agents, subagents, prompts, contexts, memory, hooks, MCP. He said it came with no manual and everyone has to figure out how to hold it. This video is part of that manual. Hooks are one of the most https://t.co/7UWilUDwUx
— Ray Fernando (@RayFernando1337) Dec 28, 2025
Absolutely incredible stuff from the Claude Code team. I used to be a regular subscriber to the highest paid versions for Claude Code, Gemini, and ChatGPT. But after last month, I am seriously considering removing the ChatGPT and Gemini subscriptions and only exclusively doing a https://t.co/umlPNVGlzw
— Soumitra Shukla (@soumitrashukla9) Dec 28, 2025
(Long, boring, enterprise-y note since I've been thinking about this the last couple of days) Maybe just me but - if you've spent enough time working in a large enterprise on cross-functional areas like deal desk or revops (or even the much hated compliance team), you know that https://t.co/soMH4jyZ0I
— Aarthi Ramamurthy (@aarthir) Dec 28, 2025
Lots of buzz on Claude Code here today. This short course, created with Anthropic, is the best way to learn to use it well. Please enjoy it! https://t.co/DvEZ8NWq12
— Andrew Ng (@AndrewYNg) Dec 29, 2025
Translation: we are about to accelerate because of AI breakthroughs like world models. https://t.co/1Qi3mrpovc
— Robert Scoble (@Scobleizer) Dec 28, 2025
Nice intro into Claude Code (and Codex) I am currently also using both in my workflows (Opus 4.5 & GPT-5.2 Codex) https://t.co/gMHXRa65NG
— Tibor Blaho (@btibor91) Dec 28, 2025
many people don’t know that Claude Code started as a personal side project that spread in the internal Anthropic team like a virus. this is a perfect article on how Claude Code was built, its tech stack, and how it’s being developed. highly recommended: read here https://t.co/4ePDV4N3fB https://t.co/7b7pxyl7S1
— ℏεsam (@Hesamation) Dec 28, 2025
Part 2: https://t.co/q1ao9I953D
— Jaya Gupta (@JayaGup10) Dec 28, 2025
The impressive part is that Claude figured out HOW to control them with zero prior knowledge of the Lutron system. It scanned local networks, identified open ports, pulled firmware metadata, searched the web for PDFs, and decoded the authentication flow. Claude treated https://t.co/0D63E3tuBT
— Aakash Gupta (@aakashg0) Dec 28, 2025
People have been dropping these really thoughtful X articles on the future of AI. Been a goldmine of good ideas. Here - Aaron argues that AI agents slash task costs for everyone, creating competitive parity where small teams get access to same leverage as large organizations. https://t.co/2NjpPiEtB2 https://t.co/JEzD2FuXfC
— Shanu Mathew (@ShanuMathew93) Dec 27, 2025
🤌🏻 https://t.co/9QQlZcOeFW
— Avinier (@avinierx) Dec 27, 2025
every PM building AI products should watch this (1hr 20min) Aman Khan (AI PM at Arize, ex-Cruise self-driving, ex-Spotify ML) shares the complete evaluation framework for shipping AI that actually works. I distilled my main notes, but watch the whole video: https://t.co/xRE7xFGpFw
— George from 🕹prodmgmt.world (@nurijanian) Dec 27, 2025
History will show that the AI bubble talk and circularity fears from media elites and pundits came right as the AI infrastructure buildout supercycle shifted into a higher gear. because of real demand. This is DeepSeek false FUD part two. https://t.co/qjCa2EE6An https://t.co/k1Du3xQ33X
— tae kim (@firstadopter) Dec 26, 2025
To give anything less than your best is to sacrifice the gift. —Steve Prefontaine
— Garry Tan (@garrytan) Dec 26, 2025
Excellent for all founders https://t.co/34YRXeUCRT
— Martin Tobias (Pre-Seed VC) (@MartinGTobias) Dec 26, 2025
"At Manus, a feature has to be 屌 (diào)—fucking sick or it doesn't ship." @ivanleomk wrote about what it's like on our team: https://t.co/z2T5ttczfT
— Parker Lyman (@parker_lyman) Dec 26, 2025
Karpathy sending the entire internet into an existential anxiety attack before end of the year 😭😭😭 https://t.co/0itHMV5P9d
— Xiao Ma (@infoxiao) Dec 26, 2025
$NVDA $GFS Nvidia and Groq Strategic Licensing and Talent Integration Analysis Publicly confirmed information indicates that the Nvidia–Groq event disclosed on 2025-12-24 is structured as a non-exclusive inference-technology licensing agreement combined with the hiring of Groq’s https://t.co/W46Kfj8Tzd https://t.co/MtDStrQmwo
— TheValueist (@TheValueist) Dec 25, 2025
Fun fact, I wrote the first 30 pieces with Manus' help. It's not perfect yet, but now you can most of our features at a glance! Who wants my instructions and prompts 👀 1. Start a project 2. Write custom instructions 3. Add files 4. Manus drafts them in md format 5. Upload https://t.co/pAkG0ylhxd https://t.co/eCMHPJuOiL
— Natalie (@gr00vyfairy) Dec 24, 2025
AI's real superpower: consuming, not creating https://t.co/oQYR7XlhEe
— Bilgin Ibryam (@bibryam) Dec 25, 2025
Coming in Q1. Save tens of thousands of dollars on legal expenses. Built for Pre-seed to Series B companies. https://t.co/mi02rcT6WY
— Brandon Gell (@bran_don_gell) Dec 24, 2025
Most “agent products” today don’t own the decision. They generate a draft action, and then a human finalizes it somewhere else. When that happens, the system never produces a clean, machine-readable record of why the decision was made. A decision trace is the missing artifact. https://t.co/nAXpjYb0uw
— Jaya Gupta (@JayaGup10) Dec 25, 2025
“Software is eating the world, and now AI is eating software.” https://t.co/bqvzUEEkqG
— Naval (@naval) Dec 24, 2025
This is a tacit knowledge problem, reframed as an infrastructure opportunity. All of the current AI implementations miss the reasoning behind decisions, the "why" that lives in experts' heads. You cannot "invent" a digital expert persona using just prompt engineering. You have https://t.co/rynYP4qZYB https://t.co/V4bsq2CEiw
— Muratcan Koylan (@koylanai) Dec 24, 2025
The Ecosystem Advantage $GOOG $GOOGL read the full piece here: https://t.co/ka1OeBA9ac https://t.co/SsJmGCEMlF
— Mostly Borrowed Ideas (@borrowed_ideas) Dec 23, 2025
https://t.co/uPXcTUEsnc
— Jaya Gupta (@JayaGup10) Dec 23, 2025
$GOOGL's Gemini success & TPUs are key drivers for the whole company, supporting its AI, Search, Waymo, workspace, and cloud segments. In 2026, the key AI metric for businesses will be INTELLIGENCE per COST, and GCP will be the best provider of it. https://t.co/dTwZUO1mou
— Rihard Jarc (@RihardJarc) Dec 23, 2025
Amazing this video is made of AI but it is very good. You should all see it.⚡️ https://t.co/SpgpXyLLET
— Ayurveda With Shyam (@shyam_ayurveda) Dec 21, 2025
- be Manus - have two products in a16z top50 AI list - $0 - $100M ARR under 9 months - outperforming frontier model labs on key benchmarks this is the journey of how it happened (from my personal bias😎) https://t.co/rpbllkcMuv
— hidecloud (@hidecloud) Dec 22, 2025
From Karpathy’s latest post: “LLM apps will organize, finetune and actually animate teams of them into deployed professionals in specific verticals by supplying private data, sensors and actuators and feedback loops.” This is exactly right. In a world of AI agents, there is a https://t.co/sTPrT0rP4F
— Aaron Levie (@levie) Dec 21, 2025
@eldsjal The Barry Diller bio is a great listen ( he reads it himself). Lots of good bios this year. The Bill Buckley bio is maybe the best biography in a long time.
— Sriram Krishnan (@sriramk) Dec 20, 2025
Karpathy’s 2025 retrospective is the clearest articulation I’ve seen of what foundational AI labs are actually building. We’re not “evolving animals,” we’re “summoning ghosts.” LLMs have completely different optimization pressures than biological intelligence. Humans evolved https://t.co/KCimA5eNw4
— Aakash Gupta (@aakashg0) Dec 20, 2025
This is it. I’ve peaked. There’s nowhere to go from here. @simonw you’re incredibly kind, thank you so much. I’m glad you enjoyed it. https://t.co/BKD6O99sYG
— Sam Rose (@samwhoo) Dec 19, 2025
Some technologies excite me. Others wake me up. GeoSpy does both. GeoSpy is a new AI tool that can pinpoint the exact location of any photo — even when GPS data is stripped out. At first glance, it feels like magic: the system matches small visual clues in the background with https://t.co/jVrocN9qlJ
— Pascal Bornet (@pascal_bornet) Dec 19, 2025
🚨BREAKING Claude is officially browser-native. The new Claude for Chrome extension runs end-to-end workflows in Chrome 🤯 You can: → Pull dashboard data into one analysis doc → Handle slide comments automatically → Build with Claude Code, test directly in Chrome and more! https://t.co/MyMWVq8yXl
— Charly Wargnier (@DataChaz) Dec 19, 2025
"This is the highest leverage moment for potentially a single individual in the history of the world" ~ John Carmack We are so lucky to be living in this rare moment in the history of humanity. https://t.co/Y01bdb9sq3
— Rohan Paul (@rohanpaul_ai) Dec 18, 2025
Either Google or OpenAI need to buy Manus. At the very least, they should be inspired by this flow: Generate slides, edit with click and change. This is the way. https://t.co/Qn28GGrzFa
— Simon Smith (@_simonsmith) Dec 18, 2025
I published my $AMZN Trainium deep dive. Investors are underestimating this ASIC. 1. Specs & software stack 2. Performance & costs of Trainium 2 & 3 3. Cost comparison of Trainium, $NVDA, and $GOOGL TPU 4. Bottlenecks 5. AWS Trainium business strategy https://t.co/CgSO4iXoZW
— Rihard Jarc (@RihardJarc) Dec 18, 2025
You can now edit slides created with Nano Banana Pro on Manus Not regenerate. Edit. Click on text and change it. Point to a visual and adjust it. Select multiple areas and update them all at once. It's what AI presentations should have been from the start. https://t.co/noc2gQVpt3
— Manus (@ManusAI) Dec 18, 2025
$0 → $100M ARR in 8 months. Since we launched in March: -147 trillion tokens processed -80M+ virtual computers created -Total revenue run rate over $125M Thank you to everyone building with us. https://t.co/TZJ3n162zl
— Manus (@ManusAI) Dec 17, 2025
This new model lets you talk to an AI you can see, just like a video call. In two years this will be common, but for now it is a pretty magical experience to talk to an AI that looks like a person. https://t.co/TmTe0HdrO3
— Jared Friedman (@snowmaker) Dec 17, 2025
Worth reading all of https://t.co/F4bmH7nUbD https://t.co/4Zi41rbc5r
— Sriram Krishnan (@sriramk) Dec 17, 2025
The Agentic OS is here! Today, we’re launching Origon, a foundation for agentic AI, built for the era when intelligence becomes software. No cloud APIs. No frameworks. No glue code. LLMs, orchestration, parallel execution, memory, and tools, unified in a single platform https://t.co/PxIF2OL0Hh
— Origon (@origon_ai) Dec 16, 2025
the idea of retirement is so bizarre to me. it assumes the goal of life is to exit effort & replace it with leisure, as if the brain doesn’t atrophy without any resistance. for most ppl work isn’t the problem.. it’s usually *meaningless* work. the greatest gift god can give you https://t.co/7BABwco8AU
— signüll (@signulll) Dec 16, 2025
We’ve pushed out the Pareto frontier of efficiency vs. intelligence again. With Gemini 3 Flash ⚡️, we are seeing reasoning capabilities previously reserved for our largest models, now running at Flash-level latency. This opens up entirely new categories of near real-time https://t.co/q0y9wfQEgW
— Jeff Dean (@JeffDean) Dec 17, 2025
🚨BREAKING: Meta just dropped SAM Audio, and it changed audio editing forever. You can now separate any sound from any audio or video using simple text prompts. https://t.co/6Bt389drKd https://t.co/1AW0Eyi2Xc
— Poonam Soni (@CodeByPoonam) Dec 16, 2025
A new podcast featuring Demis Hassabis has been released. https://t.co/otbWWtQlgZ
— NomoreID (@Hangsiin) Dec 16, 2025
Excellent "AI eats the world" deck by @benedictevans https://t.co/bqbThoQiAH
— Boris Wertz (@bwertz) Dec 16, 2025
NEW Research from Apple. When you think about it, RAG systems are fundamentally broken. Retrieval and generation are optimized separately, retrieval selects documents based on surface-level similarity while generators produce answers without feedback about what information is https://t.co/nrSZZv2wKz
— elvis (@omarsar0) Dec 15, 2025
Sergey Brin on what Google did right and wrong with AI over past 15 years: ▫️Underinvested in productizing Transformer paper (“we didn’t take it as seriously as we should have…OpenAI ran with it.”) ▫️But made correct bets on compute with TPUs and Google Cloud (“[Other than https://t.co/I84H89OXO4
— Bearly AI (@bearlyai) Dec 15, 2025
This guy built an entire AI data science team in Python. Then open-sourced (100% free). It automates data science workflows with AI, including data loading, cleaning, exploratory analysis, and feature engineering. And it tracks each step in a 100% reproducible pipeline. 00:00 https://t.co/I4fFVdbOAk
— 🔥 Matt Dancho (Business Science) 🔥 (@mdancho84) Dec 15, 2025
Very impressive American open model launch. https://t.co/zzutZo9Jpi
— Sriram Krishnan (@sriramk) Dec 15, 2025
no way this isn't killing call centers PolyAI just got Marriott, Caesar's Palace, and Gordon Ramsay's restaurants to fire their phone staff.. this AI handles every call and you literally can't tell it's not human here's how it works: https://t.co/3hJQIhodHe
— Robin Delta (@heyrobinai) Dec 15, 2025
So, @cursor_ai was able to migrate their CMS to raw code and Markdown in a few days (instead of weeks) with Opus 4.5 Meanwhile, @OpenAI shipped their @soraofficialapp for Android in just 28 days with Codex AI won't replace people, but companies using AI will replace companies https://t.co/ZeEDSQApLO
— Niels Rogge (@NielsRogge) Dec 14, 2025
Sergey Brin on the genius of Jeff Dean. He credits Jeff’s early obsession with neural networks, back when they were “telling cats from dogs”, as the spark for everything. TPU was Jeff’s idea. He calculated that if users spoke to Google for just three minutes a day, Google would https://t.co/jdHdHJSf4g https://t.co/v1EKkYcxcj
— Yuchen Jin (@Yuchenj_UW) Dec 15, 2025
New video of Sergey Brin at Stanford University. On students quitting Computer Science due to AI: “I wouldn’t switch to comparative literature, AI can probably write that better than humans too.” --- From 'Stanford University' YT Channel (link in comment) https://t.co/9d3kDRwl0B
— Rohan Paul (@rohanpaul_ai) Dec 15, 2025
Andrej Karpathy, Eureka Labs founder & former Director of AI at Tesla, breaks down how LLMs like ChatGPT "download the Internet." Watch the full beginner-friendly breakdown here: https://t.co/doph3alGnE https://t.co/EaNowJhVdS
— MIT CSAIL (@MIT_CSAIL) Dec 14, 2025
Sergey Brin admits Google messed up by under-investing in the transformer architecture it invented Google was too scared to release chatbots that "say dumb things", so it under-invested in scaling compute "we didn't take it very seriously... and openAI ran with it" https://t.co/G6nLyGajB1
— Haider. (@slow_developer) Dec 13, 2025
Sergey on why he returned to Google: He retired in 2020 with a simple plan: sit in cafes and study physics. Then COVID hit, the cafes closed, and he felt his mind “spiraling”. So he returned to build Gemini, back in founder mode. For top minds, happiness is not leisure; it is https://t.co/PoxnLF3O63 https://t.co/v1EKkYcxcj
— Yuchen Jin (@Yuchenj_UW) Dec 14, 2025
Everyone knows Nano Banana Pro is a beast at ads, but manually prompting one image at a time is a waste of life. So I built a tool called The Ad Machine with @GoogleAIStudio & Gemini 3.0. It turns 1 product upload into endless 4K ad variations. Automatically. Auto-Prompting: https://t.co/07RZXn3KJv
— Jake Baumann (@Jake_Joseph) Nov 28, 2025
Healthcare leads vertical AI adoption Vertical AI captured $3.5B in spend in 2025 (~3x from 2024) Healthcare captures nearly half of all vertical AI spend (~$1.5B), more than tripling from $450M the year prior and exceeding the next four verticals combined. https://t.co/2sbuvhFs5O
— Healthcare AI Guy (@HealthcareAIGuy) Dec 12, 2025
Massive update for everyone building agentic applications with React. CopilotKit version 1.50 includes a new "useAgent()" universal React hook. This hook makes it much simpler to build agentic applications with React. This is now the easiest way to connect a frontend to any https://t.co/r4i7iojHPw
— Santiago (@svpino) Dec 11, 2025
Turning Claude Code into an employee part one. https://t.co/I1PiL7uBMA
— Vibecoding Explained (@learn2vibe) Dec 11, 2025
Agentic AI Foundation — advancing open-source agentic AI: https://t.co/5iGjhrMO0D
— Greg Brockman (@gdb) Dec 10, 2025
Shopify's new SimGym is AI applied right: Being able to get statistically significant A/B tests done quick without the volume of a huge store is crazy. https://t.co/EsaMclBmNn
— DHH (@dhh) Dec 10, 2025
The above insight suggests that to generate a new token, every attention operation in the network only needs: - query vector of the last token. - all key & value vectors. But, there's one more key insight here. https://t.co/jdibWuQyVd
— Avi Chawla (@_avichawla) Dec 10, 2025
Dividend stocks aren’t supposed to outperform like this. These 6 completely break that narrative: 1. Caterpillar – $CAT: +65.21% YTD https://t.co/0bIfw4ysps
— Qualtrim (@qualtrim) Dec 10, 2025
My brother sam has often said that most people don't spend nearly enough time thinking about what they should be working on. This is the venture capital version: what companies should you go deep with? "We realized a long time ago that the most important decision is actually https://t.co/We9mX4YdqK https://t.co/ODiboEVIWd
— Jack Altman (@jaltma) Dec 9, 2025
Many are saying that 2026 will be the year of the agent. A passive system of record layer stops making sense when agents can independently execute on assigned intent. I expect to see a new kind of dynamic agent layer that actually makes sense for employees replace legacy SOR. https://t.co/xlbJJAir3G https://t.co/oox2FDZVYN
— Sarah Wang (@sarahdingwang) Dec 9, 2025
Anthropic is donating the Model Context Protocol to the Agentic AI Foundation, a directed fund under the Linux Foundation. In one year, MCP has become a foundational protocol for agentic AI. Joining AAIF ensures MCP remains open and community-driven. https://t.co/718OwwyFJL
— Anthropic (@AnthropicAI) Dec 9, 2025
Reliability is the barrier holding back agents, and right now the most common way to achieve it is scaling back ambitions. https://t.co/0sgDWJdoc9
— Drew Breunig (@dbreunig) Dec 9, 2025
Another classic startup essay is @sarahtavel's "Hierarchy of Engagement". https://t.co/g79OzpD4Da Sarah is a longtime investor (most notably, Pinterest, which she also joined as a product manager) across categories, but in particular, marketplaces and consumer social. Her https://t.co/Hrhod5Kpom
— Peter Kazanjy (@Kazanjy) Dec 9, 2025
This guy literally dropped a Claude 4.5 masterclass to build full-stack apps in 493 secs https://t.co/kA5AGVe4mr
— Hayes (@hayesdev_) Dec 9, 2025
We at @meritechcapital are excited to share our 2025 markets update. Each year, we deliver a markets update presentation to our limited partners at our annual meeting and we've made it public this year. https://t.co/vGSBbLgzUj
— Alex Clayton (@afc) Dec 9, 2025
Today we hit $100M ARR at @clay. It took us six years to go from $0-1m, then two years to go from $1-100m. I’m going to walk you through the 6 biggest GTM bets that got us here. $100M ARR may be the headline, but I’m most proud of how we accomplished it: we’ve never churned https://t.co/jzyjGkt3kv
— Varun Anand (@vxanand) Dec 8, 2025
McDonald’s just dropped a new AI ad and it’s beautiful and I am genuinely tired of people pretending this is not the future of media. If this played on your TV during a normal commercial break, you would be disingenuous to say “its slop” or “I could easily tell it is AI.” It is https://t.co/E9Jdgd2sUo https://t.co/kSim2rzKvu
— Chris (@chatgpt21) Dec 9, 2025
🚀 Paper2Slides is now open source! Transform research papers & technical reports into professional presentations with ONE click! We've generated stunning presentation slides from the latest DeepSeek V3.2 paper in diverse styles - check them out and share your feedback! 🔥 Core https://t.co/8kPaRwE6Nq
— Chao Huang (@huang_chao4969) Dec 8, 2025
Last month, Google dropped something interesting: five AI Agent papers released across five consecutive days, one per day, each digging into a different part of how agents should be built, evaluated, secured, and deployed. No big splash, just a steady rollout of more than 250 https://t.co/jNrVG3l0XU
— Tech with Mak (@techNmak) Dec 9, 2025
The fastest company in history to $1B did it with no VC money and fewer than 100 people. @HelloSurgeAI has become the secret weapon behind Anthropic and Google's best models. Founder Edwin Chen (@echen) built it without playing the Silicon Valley game—no viral posts, no https://t.co/wFTMuiehau
— Lenny Rachitsky (@lennysan) Dec 7, 2025
incredible interview, dyson is goated https://t.co/L8In8kcXxJ
— Aaron Slodov (@aphysicist) Dec 8, 2025
“Intelligent Document Processing” 📑🧪 as an industry is gone . With our latest release this week, *anyone* can build and deploy a specialized document agent in seconds ⚡️🤖, and customize the steps via code. Let’s take a tour through our invoice processing and contract matching https://t.co/vRT4iC6XVh
— Jerry Liu (@jerryjliu0) Dec 7, 2025
Ok it's getting truly banana town in VC and startup land. And, I think the people who will get HURT the most in this might be the early employees. So here are some things you must absolutely know and questions for you to ask as you enter job hunting season in January.. 1)
— Nikunj Kothari (@nikunj) Dec 6, 2025
📚Google released a solid guide on context engineering for working with multi-agent systems. Shows how context engineering keeps multiagent systems fast, affordable, and less hallucination prone. The guide replaces giant prompts with a compiled view over state that splits https://t.co/2gjJzX59Fd
— Rohan Paul (@rohanpaul_ai) Dec 6, 2025
Great episode 👌 https://t.co/B452XDHcMj
— Brian Armstrong (@brian_armstrong) Dec 7, 2025
Context Engineering Approaches in Manus 1. Offload Context →Store tool results externally →Push actions to the sandbox 2. Reduce Context →Compact stale results →Summarize when needed 3. Isolate Context →Use sub-agents for discrete tasks →Share context deliberately https://t.co/g5lsz8QgYh
— Bilgin Ibryam (@bibryam) Dec 7, 2025
This is a really good report on the reality of building agents in production. https://t.co/lZlXbHfpkn
— elvis (@omarsar0) Dec 6, 2025
This is one of the better takes on the AI "bubble" I've seen recently. https://t.co/sMJsm1vnT2
— Mike Maples, Jr (@m2jr) Dec 6, 2025
Notes on building in the AI era (plus how a16z operates) https://t.co/gJmZdoBasV https://t.co/q6puXqVBCl
— a16z (@a16z) Dec 5, 2025
My top 5 fine tuning techniques are: - LoRA - LorA - LORa - LorA - LOrA https://t.co/SS0cWTXYmG
— Pau Labarta Bajo (@paulabartabajo_) Dec 5, 2025
How Netflix's acquisition of Warner Bros changes the media company pecking order More charts of the week: https://t.co/OV3JEhla7L https://t.co/nA3hYR99FB
— a16z (@a16z) Dec 5, 2025
AI models are trained on public or generally available data sources. By default, they know basically everything about anything, *other* than your specific workflows and business. For AI Agents to be effective in the enterprise, they need your enterprise context. That context is https://t.co/3UvMuJwCes
— Aaron Levie (@levie) Dec 5, 2025
This is wild. China's Alibaba just dropped Live Avatar. This AI turns any voice into a realtime, talking avatar with infinite length at 20 FPS. 10 wild demos:👇 1. Ilya interview that never happened https://t.co/Zgi5ck1ZHd
— Min Choi (@minchoi) Dec 5, 2025
How learnings & ideas from working at Goldman Sachs & Bridgewater, to Citadel & Citadel Securities, led Luana (@luanalopeslara) & Tarek (@mansourtarek_) to create Kalshi 7 years later, it's worth $11 billion. https://t.co/6W5o5l6LyG https://t.co/Yr0hvbsk5d
— Molly O’Shea (@MollySOShea) Dec 4, 2025
Antigravity is unapologetically agent first. [1/9] Welcome to the "Age of Artifacts": multi-modal agents across your IDE and browser. It's more than just a code editor. Here's what I had to say at @aiDotEngineer summit nyc a couple weeks ago 🧵 https://t.co/umds20MyCZ
— Kevin Hou (@kevinhou22) Dec 3, 2025
🚀 Introducing Agentic Context Engineering (ACE) --- a framework for self-improving language models through continuously evolving contexts (not weights). 📈 High-Performing: +10.6% on agent tasks, +8.6% on finance ⚡ Ultra-Efficient: −86.9% latency, −83.6% dollar cost 🛠 https://t.co/GIZhfucsyD
— Qizheng Zhang ✈️ NeurIPS 2025 (@qizhengz_alex) Dec 2, 2025
The multi decade dream of a high scale prediction market that helps our imperfect selves discern the future has finally arrived. https://t.co/vZwHbUJKMJ
— Marc Andreessen 🇺🇸 (@pmarca) Dec 3, 2025
New post: Thoughts on AI progress (Dec 2025) 1. What are we scaling? https://t.co/lWyVVSuwO7
— Dwarkesh Patel (@dwarkesh_sp) Dec 2, 2025
🚀 Today, we announced frontier agents, a new class of AI agents which are a step-function change in what you can do with agents. Our first three agents mark the beginning of a new era in software development: 🟠 Kiro autonomous agent maintains persistent context and learns https://t.co/xs2jNXtS4m
— Swami Sivasubramanian (@SwamiSivasubram) Dec 3, 2025
vertical AI is having a moment. Harvey, Abridge, OpenEvidence all crossed billion dollar valuations by embedding deeply into workflows that horizontal software never cracked. The playbook: own the workflow, build proprietary data moats, price on outcomes not seats. The
— Ben Smith (@bcsmithx) Dec 3, 2025
Quick Mag 7 Valuation: $NVDA - fairly valued $AAPL - overvalued $MSFT - fairly valued $GOOG - fairly valued $META - undevalued $AMZN - undervalued $TSLA - heavily overvalued Bonus: $TSM - fairly valued $ASML - fairly valued $AMD - overvalued $NFLX - overvalued Cheap Quality:
— Deep Value Investing (@DeepIceValue) Dec 3, 2025
Ben Evans released 90 slides on "AI eats the world" Here are the 14 most important ones: https://t.co/NsOMxfjgoC
— Niko Ludwig (@Collateral_com) Dec 2, 2025
How I created these landing pages with Gemini 3 from start to finish First, I start with the hero section. It includes the nav bar, eyebrow, headline, subheadline, cta, social proof and visual. I spend 50% of the time here because it sets the colors, typography, spacing, which https://t.co/W0NfCbi3SX
— Meng To (@MengTo) Dec 2, 2025
Another absolute classic startup essay is @lennysan's "A Guide for Finding Product-Market Fit in B2B". https://t.co/2f6JZg8Qkd Like so many of Lenny's posts, a work of art with a boatload of original research. Just a tour de force on coming up with B2B startup ideas, https://t.co/pZjOyXldNu
— Peter Kazanjy (@Kazanjy) Dec 2, 2025
So, which chunking strategy should you actually use? There's tons of chunking techniques out there, and as always, the "best" strategy depends on your specific use case. You have to balance two competing priorities: 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹 𝗣𝗿𝗲𝗰𝗶𝘀𝗶𝗼𝗻: Chunks need to be https://t.co/TX306bPdE0
— Victoria Slocum (@victorialslocum) Dec 2, 2025
A young man in Chengdu, China used technology to make “flying swords” in real life. He controls many drone-like swords with hand gestures, making it look like scenes from a fantasy movie. People online are amazed and call it “modern-day kung fu tech.” https://t.co/qhQyZJL2Ic
— China Perspective (@China_Fact) Dec 2, 2025
Google, Nvidia, and OpenAI OpenAI and Nvidia are both under threat from Google; I like OpenAI’s chances best, but they need an advertising model to beat Google as an Aggregator. https://t.co/3Rk3smQ2WE
— Stratechery (@stratechery) Dec 1, 2025
a good writeup on prompt caching, recommended read https://t.co/C2INiA1zp5
— Ahmad (@TheAhmadOsman) Nov 30, 2025
Wow, this is absolutely the next-level solution >Open Shortcuts app >Click 'Automation' on the bottom, then the '+' in the top right >Scroll down, click 'App', select the apps you want (I did Photos and Camera) >Select 'Is Closed' and 'Run Immediately', click 'Next' >Click https://t.co/NpMeWuC66P
— Sir Barefoot (@SirBarefoot) Dec 1, 2025
This! Sometime last year I made the switch, the default is greyscale and it automatically turns to colour when I’m looking at photos and back when I close it! 100% would recommend https://t.co/i8o7B8HBgr https://t.co/g9dQcjneWl
— Vaibhav (VB) Srivastav (@reach_vb) Nov 30, 2025
Agentic AI Creation Process 📘📚 #ai #agenticai https://t.co/Ah8w5FmWF5 https://t.co/pkdriUNibU
— Krishna Agrawal (@Krishnasagrawal) Nov 30, 2025
Interview with Nikhil https://t.co/4mmIo9rcKw
— Elon Musk (@elonmusk) Nov 30, 2025
This is a must-watch documentary on AI. The Thinking Game | Full documentary | Tribeca Film Festival official se... https://t.co/kYx9glOAad via @YouTube
— Dave Kellogg (@Kellblog) Nov 30, 2025
McKinsey published new report. Agents, robots, and us: Skill partnerships in the age of AI - Today’s technologies could theoretically automate more than half of current US work hours. This reflects how profoundly work may change - By 2030, about $2.9 trillion of economic value https://t.co/kXKQjL42b9
— Rohan Paul (@rohanpaul_ai) Nov 29, 2025
https://t.co/HqxebZors7
— Jukan (@jukan05) Nov 30, 2025
It’s honestly such an honor. Jensen Huang read the Chinese article I translated, and they even shared it internally. https://t.co/xzYuAZEWAK
— Jukan (@jukan05) Nov 30, 2025
Cursor 2.1 might be the first IDE that thinks. Here’s what shocked me when I tested it: → It questions you like a product manager → It reviews code like a senior dev → It finds patterns across 500+ files instantly → It catches null bugs before you see them → It tests UI https://t.co/pozIWPBmmL
— Julian Goldie SEO (@JulianGoldieSEO) Nov 29, 2025
@fchollet While Satya was figuring out the layers, Google turned them into a dance routine. https://t.co/OrGbcYHQ1h
— Vamsi Nimmala— వంశీ🪈Γ(r,t) (@vamsi_nimmala) Nov 29, 2025
Seeing Like A Software Company is one of the best essays on legible and illegible work I’ve read. It translates Seeing Like A State (James C. Scott) into SV speak. https://t.co/0G7o9kGuEb
— Max Schoening (@mschoening) Nov 29, 2025
Everyone knows Nano Banana Pro is a beast at ads, but manually prompting one image at a time is a waste of life. So I built a tool called The Ad Machine with @GoogleAIStudio & Gemini 3.0. It turns 1 product upload into endless 4K ad variations. Automatically. Auto-Prompting: https://t.co/07RZXn3KJv
— Jake Baumann (@Jake_Joseph) Nov 28, 2025
Work life balance is great, in-fact we recommend it to all our competitors. Don’t be angry with this post. You’re free to disagree & let us cook :) https://t.co/SjzL0xs2m0
— Ayush Jaiswal (@ayushjaiswal) Nov 29, 2025
How AWS Implements Load Balancing. → AWS uses load balancing to distribute incoming traffic across multiple resources for high availability, fault tolerance, and better performance. → The main service that provides this capability is the Elastic Load Balancing (ELB) family. https://t.co/q8UlLY3VQA
— Dhanian 🗯️ (@e_opore) Nov 29, 2025
. @lukaszkaiser on what a reasoning model actually is: “A reasoning model is like your base LLM, but before giving you the answer, it thinks in the chain of thought, meaning it generates some tokens, some text that's meant not for you to read, but for the model to give you the https://t.co/MJt6XsFMsF https://t.co/turSRIFHK6
— Matt Turck (@mattturck) Nov 28, 2025
This Gemini mega prompt will help you build complete AI agents: Here's the full prompt: "You are an expert n8n workflow automation architect specializing in building AI-powered agents. I need you to design and build a complete n8n workflow for the following use case: https://t.co/Dglam2Uqst
— Brady Long (@thisguyknowsai) Nov 29, 2025
Machine Learning Lifecycle (step by step) https://t.co/HrYG61gOeH https://t.co/2juZq0PJcX
— Tech Fusionist (@techyoutbe) Nov 28, 2025
Sundar Pichai just dropped a surprisingly candid look into Google DeepMind’s current AI heartbeat. Here’s the essence of the clip: • Google is now operating on a steady six month cadence, pushing the frontier twice a year • Sundar openly admits that meaningful leaps are https://t.co/X7xqFD0yAt
— VraserX e/acc (@VraserX) Nov 28, 2025
How AI Agents Use Memory Systems 1. Introduction → Memory is essential for AI agents because it allows them to retain information, reason across time, and improve decisions based on past interactions. → Without memory, agents would act blindly, unable to learn or adapt. 2. https://t.co/U4SuhN33NB
— Dhanian 🗯️ (@e_opore) Nov 28, 2025
You’re using NotebookLM wrong. Here’s the right way to extract maximum value: 1. Create a notebook for ONE topic 2. Upload 10–50 sources 3. Use deep research to expand your library 4. Ask questions only based on sources 5. Save the best answers as notes 6. Convert https://t.co/3wrWtx43qL
— Julian Goldie SEO (@JulianGoldieSEO) Nov 28, 2025
for those of you just finding out abt TPUs, here’s a great educational resource to learn more abt them :) https://t.co/wNR9nEbknn
— surya (@suryasure05) Nov 27, 2025
🚨 This project just made most AI agents look outdated. It’s called Better Agents and it supercharges your coding assistant (Kilocode, Claude Code, Cursor, etc), making it an expert in any agent framework you choose (Agno, Mastra, etc) and all their best practices. This is the https://t.co/HgcX2PoLM4
— Hasan Toor (@hasantoxr) Nov 27, 2025
One thing I keep noticing: top performers in any role can "codify" excellence. Great example from @bobmcgrewai. https://t.co/oQy78UFx2P
— Vedika Jain (@vedikaja_in) Nov 26, 2025
Banger paper from NVIDIA. Bigger models aren't always the answer. However, the default approach to improving AI systems today remains scaling up. More parameters, more compute, more cost. But many tasks don't require the full power of a massive model. This new research https://t.co/hQFWDqSLP9
— DAIR.AI (@dair_ai) Nov 27, 2025
This book is a pretty important part of OpenAI lore. Greg Brockman’s post on Ilya in 2024. https://t.co/W9cq9N4m6j https://t.co/gGopXoY671
— Blake Robbins (@blakeir) Nov 26, 2025
New Anthropic research: Estimating AI productivity gains from Claude conversations. The Anthropic Economic Index tells us where Claude is used, and for which tasks. But it doesn’t tell us how useful Claude is. How much time does it save? https://t.co/jDscH6j6iX
— Anthropic (@AnthropicAI) Nov 25, 2025
