My mom called. Apparently the neighbor’s son just launched a context company. She wanted to know: if everyone is saying they do context now, what makes ours different? Fair question. This is my answer. 👇 On April 29, we're not just explaining the context layer. We're building one live. No slides. The real thing.

Coding agents learn from experience, but that knowledge stays locked in silos. Solve a thousand SWE tasks, and none of that wisdom helps with competitive coding. What if memories could transfer across domains? The work introduces Memory Transfer Learning, a framework where coding agents share a unified memory pool across 6 heterogeneous benchmarks. They test four memory formats ranging from raw execution traces to high-level insights, and find that cross-domain memory improves average performance by 3.7%. Why does it matter? The transferable value isn't task-specific code. It's meta-knowledge: validation routines, structured action workflows, safe interaction patterns with execution environments. Algorithmic strategy transfer accounts for only 5.5% of the gains. The real benefit comes from procedural guidance on how to act, not what to code. Abstraction dictates transferability: high-level insights generalize well, while low-level execution traces often cause negative transfer by anchoring agents to incompatible implementation details. Paper: https://t.co/XPD5kczsoZ Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

@NousResearch The Agent just leveled up again! gaining new skills at a steady rate. Noice! https://t.co/KnS3yD2lKL

Disturbing FACT: 30% of Gen Z Has Been Aborted That statistic blows my mind away... https://t.co/DkSWhWJelC

Disturbing FACT: 30% of Gen Z Has Been Aborted That statistic blows my mind away... https://t.co/DkSWhWJelC

GALLUP POLL: 42% of men aged 18-29 now say religion is "very important" in their lives — a sharp jump from just 28% in 2022-2023. Monthly religious attendance among young men has climbed to 40% (up from 33%), the highest level in over a decade. https://t.co/3lMO2Y6kFm https://t.co/np8SLotoFe

80% of outcomes in life can be explained by 1 thing: Pareto principle. The pareto principle is the pareto for principles. https://t.co/eRbC5zwN5W

这是Hermes也下场玩中转站了吗？还集成到了自家产品 不一样的是，一次订阅能除了300+模型还有第三方付费tool，各家抓网页、生图、语音月费从10到100刀 @Teknium @NousResearch this is definitely a good business idea 👍 #hermes #中转站 https://t.co/3eptijauem

@fujikanaeda Having 0 issues https://t.co/xpBTXVNpnV

The new Codex is another jump in what agents will look like for knowledge workers. Agents that can code, work with tools, and use computers, can begin to execute long running tasks in the background for all areas of work. This can mean drafting reports, setting up data rooms for a merger, reviewing contracts, helping onboard clients, generating marketing assets, processing invoices, and more. With the Box plugin inside of the new Codex, you can begin to automate almost any kind of work with enterprise content. And importantly, being able to work across multiple apps is a huge point of leverage because we can now far more easily connect our tools together.

Big day for Codex! Codex can now work across more of your computer and more of your tools. Features many of you have been asking for are here: computer use, in-app browser, 90+ plugins, image generation, memory, thread automations, and more. Can’t wait to see what you build! https://t.co/TD2plrlHpM

Ray Kurzweil dropped a mind-bending clarification on the timeline we’ve all been watching. He still stands by his famous 2029 prediction for human-level AI. But the Singularity — the point where we become a thousand times smarter — comes in 2045. The key difference? We won’t just have AI beside us. We’re going to merge with it. Kurzweil says the boundary will disappear: you won’t know whether an idea popped into your head from your biological brain or your computational intelligence. It will feel exactly the same. Today we can tell when something comes from an LLM. In the future, we won’t. By 2045, he believes this merger will multiply our intelligence by a factor of 1,000. It’s not two separate things competing. It’s one seamless intelligence — and we become it. What do you think — does the idea of fully merging with AI excite you, scare you, or both?

We do!! @SophontAI has released the Medmarks benchmark suite, which is the largest completely open-source automated evaluation suite for medical capabilities. (new version coming soon) We'd love to help any frontier lab evaluate their model using our suite! https://t.co/ACNe1b9Vko

Check out the new Codex Plugins, you can find us there! Within your Codex app you can now install the Remotion Skill with one click, skipping the CLI installation. https://t.co/kazdgkdv6w https://t.co/xr6THCy42V

Anthropic says Opus 4.7 hits 80.6% on Document Reasoning — up from 57.1%. But "reasoning about documents" ≠ "parsing documents for agents." We ran it on ParseBench. → Charts: 13.5% → 55.8% (+42.3) — huge → Formatting: 64.2% → 69.4% (+5.2) → Content: 89.7% → 90.3% (+0.6) → Tables: 86.5% → 87.2% (+0.7) → Layout: 16.5% → 14.0% (-2.5) — regressed Real chart gains, but at ~1.5¢/page. Enterprise scale? Not yet. LlamaParse Agentic: 84.9% overall. ~1.2¢/page. The frontier for general document understanding is long. No single model solves it. → https://t.co/h7SpuTWYVn

Tool Gateway is now live in Nous Portal. No separate accounts, no API key juggling. All you need is one subscription, and everything works. A paid Nous Portal subscription now includes access to 300+ models and a growing set of third-party tools. Launching with: → Web scraping → Browser automation → Image generation → Cloud terminal backend → Text-to-speech

On the plus side with Opus 4.7, if it does decide to think it produces BY FAR the best Sparks unicorn* ever, even non-thinking is pretty good, if not great. * This is created using TikZ, which is a language built for scientific diagrams & very much not for drawing. The original "Sparks of AGI" paper used the ability of the AI to draw a primitive unicorn as an example of unexpected AI abilities

who made this 🇫🇷 banger lol (identify yourself) https://t.co/gkfTtdR1w2

who made this 🇫🇷 banger lol (identify yourself) https://t.co/gkfTtdR1w2

You deserve more than a crowded Blue Bottle and a batch of 400 startups... I'm launching Kernel Grants, a pre-seed program that will invest $271,828 in 10 founders/year who are building tooling and infrastructure for the token factories of the future. We have an amazing set of speakers lined up for our first set of events: - @pirroh, President of Replit - @soumithchintala, CTO of Thinky - @jeremyphoward, Founder of https://t.co/RHjK7ZPIFM - @NaderLikeLadder, Dir. of DevTech at NVIDIA - @OfficialLoganK, MOTS at DeepMind (and first ever Latent Space guest!) - @clattner_llvm, Founder of Modular - @dylan522p, Founder of SemiAnalysis - @swyx, Editor of Latent Space (+ AIE, Cognition, etc!) Batches are a relic of pre-AI acceleration. Any day is a great day to start building, so applications are open and we accept founders on a rolling basis. Let's build! https://t.co/KM8KtRiniH Enjoy an exclusive tour of our Kernel space 👀

+1 more

karpathy said let there be descent credit: @rxdyxn https://t.co/Py8aO018H9

karpathy said let there be descent credit: @rxdyxn https://t.co/Py8aO018H9

Tell all the truth but tell it slant— Success in Circuit lies Too bright for our infirm Delight The Truth's superb surprise This paper finds poetry is a universal single shot jailbreak for LLMs. Systems built to stop prosaic attacks fail when the request is phrased in verse. https://t.co/uekU0l9QdL

Introducing Arrow 1.1 and Arrow 1.1 Max Our most advanced and capable models for structured vector generation Read more ↓ https://t.co/awYauwjYac

We are super excited to launch the in-app browser inside Codex with comment mode! View any web pages & iterate with your agent quickly with just point and click. Codex will automatically capture a screenshot, the DOM element, and feed it as precise context to your next chat. No more switching between browsers, dragging screenshots, and wrangling with underspecified prompts. It's great for front-end development of apps/pages, but also very useful if you have documentation pulled up on the side and just want to ask a question!

It basically rarely seems to think on analysis, writing, or research tasks, which means it isn't using tools or web search. Haven't tested everything yet, so not definitive, but I am often getting lower quality answers for that sort of use case that Opus 4.6 Extended Thinking. https://t.co/svmp46CkTJ

5 steps to making codex your chief of staff 1. download the desktop app 2. install the plugins you need for work 3. paste this into a thread and pin it 4. ??? 5. monitor the situation https://t.co/6ioazy9xvb

I have found that asking for a sestina regularly triggers Opus 4.7's safety guardrails. The forbidden poetic form! https://t.co/5Jnfhx0Qff

I think the adaptive thinking requirement in Claude Opus 4.7 is bad in the ways that all AI effort routers are bad, but magnified by the fact that there is no manual override like in ChatGPT. It regularly decides that non-math/code stuff is "low effort" & produces worse results. https://t.co/OEMM6TUpOL

Publishers have real questions about AI, but let’s be clear: @waybackmachine isn’t a backdoor for AI scraping. For 30 years, it’s been built for people, not bulk harvesting. We actively monitor to prevent abuse. Learn more ⤵️ https://t.co/YKDkawYd5G

Daniel Moreno-Gama, in an interview before he arrived in SF with a gun and a hit list: https://t.co/0LFx3QREfq

Daniel Moreno-Gama, in an interview before he arrived in SF with a gun and a hit list: https://t.co/0LFx3QREfq