Knowledge catalogue
A belated Mother's Day spotlight on a builder worth celebrating. Sarra is a technical writer and mom in Montréal who came to Replit with one goal: turn screen time into something her kids could create
AI powered creativity on full display. Vibecon transforms NYC into a living gallery of code-driven installations, June 17–18. Two days where code becomes the environment and the medium becomes the mes
This article likely discusses how advances in large language models, prompt engineering, and in-context learning are making traditional finetuning less necessary for many applications. It probably exp
This is a casual thank-you post from Swyx acknowledging @teej_m for reading his content on X (formerly Twitter). The post appears to be a brief, informal social media acknowledgment without substantiv
Computer is secure by default. Every task runs in its own hardware-isolated sandbox with VPC-level storage and compute separation. Agents are authenticated with short-lived proxy tokens instead of raw
Each environment now has its own version history with rollback. An audit log captures every change. Egress and secrets can be scoped per environment. Secrets in one environment aren't accessible from
External content is scanned in parallel by ML classifiers and the BrowseSafe model before agents act on it. File connector data is encrypted in transit and at rest, uploaded files automatically delete
if your reaction to this is “haha openclaw bad, see prompt injection is the #1 danger” you: 1) havent sufficiently appreciated the layers to this tweet 2) havent seen enough ai api keys @gilpinskyy @d
This post outlines a framework for progressively automating AI workflows through three levels of increasing autonomy: using preset prompts for consistency, refining inputs with human judgment to impro
Starting today, you can run cloud agents inside fully configured development environments. Set them up the same way you'd set up a laptop for an engineer: cloned repos, installed dependencies, and too
The UK AISI found Mythos Preview is the first model to solve both their cyber ranges end-to-end. No model had ever solved the AISI’s “Cooling Tower” cyber range before. We're getting it to defenders a
Vercel introduced a trusted sources feature for deployment protection that allows teams to control which sources can trigger deployments, enhancing security by restricting deployment access to verifie
🔴 We're LIVE from the SaaStr floor, Day 2! 🎙️🚀 Joined by an amazing lineup today: • Lindsay Wise + Julia Holm (Green Security) • Kody Low (Replit Field Engineering) • Philipp Dietz (Replit Brand & Cre
We’re training models wrong and it’s due to chatGPT. Even the modern coding agents used daily still use message-based exchanges: They send messages to users, to themselves (CoT) and to tools, and rece
Thinking Machines released TML-Interaction-Small, a 276B parameter model with a 12B active subset designed for real-time voice interactions that reportedly surpasses standard voice activity detection
arXiv:2604.08178v2 Announce Type: replace Abstract: In classical Reinforcement Learning from Human Feedback (RLHF), Reward Models (RMs) serve as the fundamental signal provider for model alignment. As
arXiv:2605.08891v1 Announce Type: new Abstract: Sparse autoencoders have become a standard tool for uncovering interpretable latent representations in neural networks. Yet salient concepts often span
arXiv:2603.25777v2 Announce Type: replace-cross Abstract: There is great potential for the application of AI tools in fusion research, and substantial worldwide benefit if fusion power is realised. Ho
arXiv:2605.10005v1 Announce Type: cross Abstract: In this paper, we describe early work on a specification inference tool for the Move Prover that combines a weakest-precondition (WP) analysis over Mo
arXiv:2508.04660v2 Announce Type: replace Abstract: Group Relative Policy Optimization (GRPO) has proven to be an effective tool for post-training language models (LMs). However, AI systems are increa
arXiv:2605.09998v1 Announce Type: cross Abstract: Coding harnesses such as Claude Code and OpenHands wrap foundation models with tools, memory, and planning, but no equivalent exists for embodied agen
Vercel introduced a natural language interface for creating Web Application Firewall (WAF) custom rules, allowing users to define firewall policies without writing complex rule syntax. This feature si
Windsurf is an AI-powered code editor developed by Codeium that combines traditional IDE functionality with generative AI capabilities for software development. The post promotes downloading Windsurf
Vercel has introduced a fast mode option for Claude Opus 4.7 on its AI Gateway service, enabling faster response times for users integrating the model through their platform. This feature likely provi
Replit announced a live broadcast or streaming event from SaaStr, likely a major software-as-a-service conference. The post suggests Replit was presenting, demonstrating, or participating in real-time
This post references an episode of Unsupervised Learning (likely a podcast or show) from 2026 that inspired a related article, featuring hosts or contributors Swyx and Jacob Effron discussing developm
https://x.com/chhillee/status/2053940218747842619?s=46 In modern ML accelerators, FLOPS have absolutely exploded. Often though, the bottleneck is not FLOPS but memory bandwidth. Similarly, model intel
This was a live broadcast from the SaaStr conference kickoff event on Day 1, streamed by Replit on X (formerly Twitter). The stream likely featured keynote presentations, panel discussions, or network
This Vercel changelog entry announces the ability to manage Vercel's firewall settings directly through the command-line interface (CLI), enabling developers to configure and control firewall rules wi
my highlights https://x.com/latentspacepod/status/2054057835043139834?s=20 [AINews] TML-Interaction-Small 276B-A12B https://www.latent.space/p/ainews-thinking-machines-native-interaction Thinking Mach
Vercel has made Node.js 26.x available for use in Vercel Sandboxes, enabling developers to test and run applications with the latest Node.js runtime version. This update allows users to leverage new f
arXiv:2605.09055v1 Announce Type: cross Abstract: Recent agentic-robotics systems, from Code-asPolicies to modern vision-language-action (VLA) foundation models, presuppose that drivers, SDKs, or ROS-
arXiv:2603.10165v2 Announce Type: replace-cross Abstract: Every agent interaction generates a next-state signal, namely the user reply, tool output, terminal or GUI state change that follows each acti
arXiv:2605.08904v1 Announce Type: new Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities in reasoning and tool use. However, the fundamental cognitive faculties essential
arXiv:2509.21677v2 Announce Type: replace Abstract: We present Prophecy, a tool for automatically inferring formal properties of feed-forward neural networks. Prophecy is based on the observation that
The thing about 90% of TDMs [Technical Decision Makers] is that they're motivated primarily by NOT GETTING FIRED. These aren't people who browser Lobsters or push to GH on the weekend. These are peopl
Replit is going to London ✈️ @posthog CEO, @james406, and @amasad are coming together for a fireside chat, hosted with our friends at @meetgranola. May 21st. Save your spot ⠕ https://luma.com/amjad-ja
The EDA Primer covers the semiconductor design and manufacturing workflow, explaining how Electronic Design Automation tools transform Register Transfer Level (RTL) code into physical silicon through
arXiv:2605.10038v1 Announce Type: new Abstract: Time series analysis underpins forecasting, monitoring, and decision making in domains such as finance and weather, where solving a task often requires
We're live from the floor at SaaStr today! 🎙️🚀 Joined by some amazing guests straight from the SaaStr stage: Amjad Masad (@amasad), Michele Catasta (@pirroh), and Ghazi Masood. Talking about how B2B c
arXiv:2605.08563v1 Announce Type: new Abstract: When an LLM agent fails a multi-step tool-augmented task and retries, the failed attempt typically remains in its context window -- contaminating the ne
Wrote about today's GitLab restructuring / 'workforce reduction' announcement, and ended up digging around in version control for both the GitLab and the 37signals public employee handbooks to help il
yeah this is the year https://x.com/ivanhzhao/status/2054273186699636975?s=46 Closed Q1 last week. Revenue is accelerating for the 7th straight quarter. AI now counts for 60% of our business. Cash flo
𝐅𝐮𝐥𝐥-𝐏𝐚𝐫𝐚𝐦 𝐑𝐋 𝐧𝐨𝐰 𝐚𝐯𝐚𝐢𝐥𝐚𝐛𝐥𝐞 𝐟𝐨𝐫 𝐊𝐢𝐦𝐢 𝐊𝟐.𝟔 You've been told only 3 AI labs matter. The best AI apps never believed that. @cursor_ai, @vercel, @genspark_ai don't run only off-the-shelf models. They trai
'Race to Revenue' is Replit's competition series tracking 20 builders competing over 8 weeks to earn their first dollar through shipped products. Episode 3 features updates on participants' progress t
Vercel Flags enables automated progressive rollouts, allowing developers to gradually deploy features to increasing percentages of users rather than releasing to everyone simultaneously. This approach
arXiv:2605.06997v1 Announce Type: new Abstract: Long chain-of-thought reasoning and agentic tool-calling produce traces spanning tens of thousands of tokens, yet Transformer KV caches grow linearly wi
For browser-use AI agents, every task is dozens of model calls in a tight loop. The inference layer isn’t background infrastructure. It’s what the product runs on. @yutori_ai runs Scouts, Delegate, an
As agents start changing software, they need a way to verify their work that includes traces, evals, feedback, and APIs. This is where Phoenix goes next — not the next release, but what this product b
I believe the kids call this '@thinkymachines just brutally framemogged gdm and oai'. basically everyone's definition of 'realtime' just got a massive frciking upgrade Media lowkey the funniest videos
I recently joined @latentspacepod to talk about AI for physics. We dug into recent work on scattering amplitudes with GPT, and what it suggests about how AI will accelerate theoretical discovery in a
Kimi K2.6 was a hit during the launch promotion - so it's staying free for Pro, Teams, and Max users for another month. Enjoy! Kimi K2.6 is now available in Windsurf! Available for free for the next 2
lowkey the funniest videos of the batch. thinky has some comedians!! congrats to @thinkymachines on reviving the omnimodel dream that others could not Today we're sharing our work on interaction model
arXiv:2605.07242v1 Announce Type: new Abstract: Agentic memory evolves across tasks into durable derived artifacts: summaries, cached outputs, embeddings, learned skills, and executable tool procedure
Swyx shared a talk that was presented at an AI Engineer event, where the speaker discussed their research and the future direction of BFL ML, with credit given to Swyx for the invitation. The full tal
openai's new $10b forward deployed company is so locked in they can't even bother to server side render properly (jk congrats, nice win for UK AI 🇬🇧) We’ve also agreed to acquire Tomoro, which will br
Qdrant 1.18.0 is out! Let’s look at the main features for this version: TurboQuant: A new quantization method that, at twice the compression ratio of scalar quantization, delivers similar recall and s
arXiv:2605.07507v1 Announce Type: new Abstract: The exponential growth of academic publications has created an urgent need for automated tools capable of extracting structured knowledge from unstructu
arXiv:2605.07237v1 Announce Type: new Abstract: Tool-integrated reasoning (TIR) has emerged as a dominant paradigm for mathematical problem solving in language models, combining natural language (NL)
the inside story of the legendary Cog House. i believe there have not been any public photos of this place until now (bc we were explicitly not allowed to lol) as an advisor its been awe inspiring to