Pratham✓
@Prathkum
I talk about web, AI, API, and social • Building experiences at @APILayer • Prev @Rapid_API @HyperspaceAI
994 Following449K Followers37K Posts
- The trust problem with AI assistants is getting worse. The ones useful enough to matter want your email, messages and files, and most copy them to the cloud. Sentient goes the other way: on-device, open source, and a knowledge base that stays on your machine as plain files. Sidekick uses its own cursor in your real apps, Double Tap writes in your voice, and a cloud model is only involved if you choose one. Easy to see why this is getting attention.Aditya Vellanki @aditya_vellankiToday, we're launching Sentient OS (YC F26), on-device AI that understands your life overnight with a local LLM. This unlimited inference enables 2 magical experiences 🧵3055.7K
Today, we're launching Sentient OS (YC F26), on-device AI that understands your life overnight with a local LLM. This unlimited inference enables 2 magical experiences 🧵8631332403K- Holy shit! This is definitely one of the best uses of Claude. Anthropic is putting Claude inside real drug development programs. Real biotech teams and real lab results. The results are wild: – Enveda: Claude reviewed 1,061 pages of manufacturing docs. It caught all 11 issues the humans found and flagged 7 they missed. Review time went from 25+ hours to under 3. A regulatory draft that takes about 3 months by hand took 4 hours. – Divergence Labs: Claude designed around 35,000 antibodies in about a month for a new food allergy drug. 20 designs pulled IgE off immune cells, and 3 looked better than the standard medication. – Generate Biomedicines: Claude Code proposed a CAR T design 70% shorter than the standard one. In lab tests, it gave roughly 10x more receptors and killed tumor cells more efficiently. – Anthropic acquired a one-person biotech startup. The founder used Claude to run the work across 11 outside labs and went from idea to results in mice in under 18 months. Helping cure disease is exactly what AI should be for.Anthropic @AnthropicAIOver the past few months, Anthropic scientists have engaged in deep research partnerships with six organizations developing new medicines. We've also begun drug discovery efforts of our own. In doing so, we’ve learned much more about how Claude can support the many stages of drug development. In this blog post, we share some early results from these efforts, and describe how this work fits in with our wider ambitions in the life sciences.171933846.6K
Over the past few months, Anthropic scientists have engaged in deep research partnerships with six organizations developing new medicines. We've also begun drug discovery efforts of our own. In doing so, we’ve learned much more about how Claude can support the many stages of drug development. In this blog post, we share some early results from these efforts, and describe how this work fits in with our wider ambitions in the life sciences.4657037.1K1.4M- Every job is becoming about managing agents, and it comes down to who can manage the most of them.2567810.8K
- Two things here: 1. At least this is better than straight up layoffs. 2. People working at bloated companies are already at a much higher risk of getting laid off anyway (I’m sorry to say this).TIMES NOW @TimesNowMicrosoft is offering voluntary exit packages to about 60,000 employees That’s roughly 27% of its workforce across the US, Canada, India and the UK Staff must decide in December, and those who accept leave on Dec. 31515610.1K
Microsoft is offering voluntary exit packages to about 60,000 employees That’s roughly 27% of its workforce across the US, Canada, India and the UK Staff must decide in December, and those who accept leave on Dec. 31531911.1K165K- OpenAI must have some strong internal models. They shared three discoveries: – Local smoothing conjecture is solved in every dimension. Humans took 28 years to crack just 2D (2019), and the related Kakeya conjecture only fell in 3D in 2025. The model solved it in every dimension. – The Exponential Time Hypothesis is false. The people who proposed it have studied it for 30 years. It was widely believed to be true and hundreds of papers were built on it. The model disproved it on Oct 16 and Lean verified it the next day. – Quantum chromodynamics (QCD) proven to exist mathematically. Physicists have used QCD for over 50 years but no one could prove the math behind it actually holds up. It also covers Yang-Mills, a Millennium Prize problem. The model did Yang-Mills on Sep 9 and QCD on Sep 18.Greg Brockman @gdbAnnouncing three major scientific discoveries, one each in mathematics (harmonic analysis), computer science (complexity theory), and physics (quantum field theory). AI starting to help us improve our understanding of the universe!112369.2K
Announcing three major scientific discoveries, one each in mathematics (harmonic analysis), computer science (complexity theory), and physics (quantum field theory). AI starting to help us improve our understanding of the universe!6472K15.6K2.9M- GPT-6.1 Terra costs $1 in / $5 out per 1M tokens. Kimi K3.1 Pro is $2 / $10. Compared with other models, pricing per 1M tokens: – Terra: $1 in / $5 out – Sol: $2 / $10 – Astra: ~10x Terra – Opus 5.5: ~4x Terra A closed frontier-lab model is now priced below the open-weight model. It’s not the best on the benchmarks. OpenAI says so itself. But most of our work doesn’t need the smartest model. Used well, Terra gets you close to frontier results.OpenAI @OpenAIGPT-6.1 Terra: delivering fast and capable computer use at a lower price. GPT-6.1 Terra understands what’s on screen, takes action across apps, writes naturally, and is a more capable model for your whole workday with the efficiency to take on more of it.814614K
- Anthropic just filed its S-1. Ticker: ANT on Nasdaq. Why it could be the biggest IPO ever: • Reported raise: $75B to $100B (SpaceX record: $75B) • ~$2T valuation talk • Run-rate revenue: $9B to $65B in 7 months • First operating profit in Q2 • OpenAI isn't listing till 2027Anthropic @AnthropicAIAnthropic has filed an S-1 registration statement with the SEC related to its proposed IPO. Read more: Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.71177.4K
- Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO. Read more: Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.1.3K2.3K18K8.3M
- One of the main reasons behind Anthropic’s success is Claude Code. And one of the main reasons behind Claude Code’s success is the incredible team behind it.Thariq @trq212I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined. I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top. The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever. My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply. To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had. For me, it is time for a bit of a break before diving into something new. If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.8114414.7K
I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined. I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top. The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever. My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply. To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had. For me, it is time for a bit of a break before diving into something new. If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.2.2K1.1K31.6K6.4M- Open weight models are defeating frontier models. Kimi K3.1 Pro is a 3T model with 1M context. I am intrigued to try it out. Scored 65.7% on Terminal-Bench 4.0. Compared to 58.2% and 66.4% for GPT-6 Astra and Opus 5.5. Open weights land by Nov 4.Kimi.ai @Kimi_MoonshotIntroducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: 🔗 Tech blog:40266.7K
Kimi.ai@Kimi_MoonshotIntroducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: 🔗 Tech blog:
1.3K3.2K30.2K4M
- This might be the most underrated thing AI has done so far. Anthropic just had Claude formally verify 123 open-source libraries at the assembly level. → 239 entry points proven crash-free for any input up to 64 MB → Bugs found that survived trillions of runs → On a PDF reader, it found a year’s worth of security bugs in one week Claude did this in a few weeks. Models are getting smarter, and hence more cyber attacks are happening across most software we thought was hard to break. We need more advancement in software verification.Anthropic @AnthropicAINew on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here:945311.6K
- New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here:6611.1K9.8K2.1M
- Wrote a piece on building apps for both humans and agents by defining the function once. Schema, validation and permissions live in the action, and every surface (UI, agent, HTTP, MCP, CLI) calls it.7485.2K
- One of my favorite companies is @Netlify. Their products are great, always head to head with the race and what the market demands, and they never create any unnecessary hype. It feels like I have been using Netlify for ages now.Mayank Gupta🧑🏻💻 @waghnakh_21had a great convo with the cto of @netlify, Dana Lawson30186.9K
- Just to summarize what happened so you don’t feel overwhelmed: – PHP2Shell Two memory-corruption bugs in PHP core give unauthenticated remote code execution (CVSS v4 9.2). They affect every version of the last decade, back to 5.4. The embargo leaked, attacks began before the patch, and govt sites went offline, including the UK NCSC. The NSA says patch or pull PHP systems offline. – Railway breach Railway detected unauthorized access to the systems that store customer environment variables. Attackers may have decrypted some of them: API keys, database credentials and private registry credentials. – LiteLLM supply chain attack Stolen maintainer tokens were used to push a trojanized release to PyPI, Docker Hub and GHCR. It steals cloud, SSH and AI keys and installs a backdoor. – Node.js Four are high severity: a Brotli decoder RCE, two TLS hostname-validation bypasses, and a dns.resolve NUL-character bug. It also distrusts the compromised eMudhra CA. Fixed in 22.23.4, 24.22.1 and 26.12.1. – SBI data exposure State Bank of India found 3,897 customer records on the dark web while reviewing about 3.75 lakh. They include names, phone numbers, DOB, addresses and loan metadata. – OpenCode Users' saved OpenAI, Anthropic, and Google API keys were compromised, and 1.4% of Console users were told to revoke and replace them. Some users report unauthorized bills from $168 to $1,700. + a lot more...1235812.7K
- Win win for both OpenAI and Google. OpenAI models get the distribution. Google gets the best models because Gemini is a little behind in the race.OpenAI @OpenAIOpenAI models are coming to Google Cloud. We're bringing our frontier models and Codex agentic workflows to Google Cloud's Model Garden, so enterprises can use our technology where they already build.9111314K
- OpenAI models are coming to Google Cloud. We're bringing our frontier models and Codex agentic workflows to Google Cloud's Model Garden, so enterprises can use our technology where they already build.4824145.3K528K
- Interesting line in Anthropic’s Report. I have seen this pattern before. – Google set out to organize the world’s information – Facebook set out to connect people – Amazon set out to be the most customer-centric company Anthropic is being honest here.
Anthropic @AnthropicAIAnthropic is a Public Benefit Corporation. Today, we’re publishing our annual Public Benefit Report. This report shares our five pillars and takes stock of the actions we’re taking to advance them. Read the full report at this page:33277.4K - Anthropic is a Public Benefit Corporation. Today, we’re publishing our annual Public Benefit Report. This report shares our five pillars and takes stock of the actions we’re taking to advance them. Read the full report at this page:2752242.6K371K
- My timeline has become a loop of three things: 1. startups raising millions 2. new AI models dropping every other day 3. companies shipping products at an unprecedented pace AI has changed the entire startup economy.1337512.9K
- Show more posts
