Knowledge catalogue
Allbirds, the sustainable footwear company known for its wool and eucalyptus-based sneakers, has reportedly rebranded or pivoted to position itself as an AI compute company, reflecting a broader trend
An undercover BBC investigation has found a ‘shadow industry of law firms and advisers is charging thousands of pounds to help migrants pretend to be gay in order to stay in the UK’. WE KNOW. We’ve al
Elon Musk posted a thank-you message on X to Taiwan Semiconductor Manufacturing Company (TSMC) and Samsung for their manufacturing support in producing a new AI chip. He indicated the chip would becom
ARC-AGI-3 is a new benchmark iteration designed so that any intelligent human applying genuine effort should achieve greater than 90% accuracy, maintaining the benchmark's core principle that tasks mu
ARC-AGI-3 has the lowest human bar of any AI benchmark out there. Almost all benchmarks require specialized knowledge that make them inaccessible to 99%+ of humans (like, say SWE-Bench). ARC-AGI-3 is
As someone who teaches at a business school, I can tell you that the pre-professional student population (as opposed to liberal arts or those pursuing academia or a topic of personal excitement) is ex
Banger paper from NVIDIA. Agentic reasoning needs models that are not just capable, but efficient at long-context inference. The agent model layer is moving toward open, long-context, high-throughput
Been waiting a month for Anthropic to answer a simple usage question about Claude Code subscriptions Have I been ghosted Can I get some questions answered by someone at Anthropic? 1. Can you use an OA
Best part are retrieval and search details: >Agent queries differ from human queries & what good results means changes too >Parallel queries and ranking are both tools to the same outcome >Top-K preci
BREAKING: DHS confirms to @FoxNews that Olaolukitan Adon Abel, the suspect arrested for a seemingly random murder spree in DeKalb County, GA that left two women dead and a homeless man shot, is a nati
btw the famous slack chart is slack propaganda and everyone who cites it is legally obligated to also link to @sophiebits Every time I see a tweet saying “I can vibe code this in a weekend” - I think
Budget blown on closed APIs is a solvable problem. Simply replace 20m of closed-source tokens with 1m on Minimax M2.7. Frontier performance. 20x lower cost. No rate limits. If you're rethinking your A
Mistral AI has launched its Connectors API into Public Preview, enabling developers to register Model Context Protocol (MCP) connectors once and deploy them across multiple Mistral products. This 'bui
Elon Musk posted on X (formerly Twitter) about Burger King advertising on the platform, likely highlighting the fast food chain's return to or continued presence on X as an advertiser. The post may re
communication is hard! structured outputs make its more clear how subagents should communicate back to the main agent deepagents now supports structured output for subagents! an under appreciated piec
Comparing hardware on InferenceX just got a massive upgrade. Our new changelog update lets you compare every change for any gpu all on the same chart. Just select any model and GPU config, then select
Compute constraints are a double bind: On the inference side you need to either (a) raise prices, (b) ration use, and/or (c) serve worse models. This hurts current growth On the training side, you can
Tesla's AI chip design team has successfully completed the tape-out of their AI5 chip, marking a significant milestone in the company's in-house semiconductor development. Elon Musk also confirmed tha
crons, part of langsmith deployments! https://docs.langchain.com/langsmith/cron-jobs I’m not sure whether LangChain supports this yet, but it might be worth considering scheduling and cron-style jobs
Cursor has introduced a new feature that allows its AI to respond by generating interactive canvases, enabling users to visually represent information beyond standard text responses. Users can prompt
Nous Research's Hermes Agent demonstrates advanced capabilities in generating digital art, showcasing how AI agents can autonomously create or assist in producing visual artwork at a sophisticated lev
Dedicated Read Nodes are now generally available. Predictable performance at scale. Up to 97% lower costs on real production workloads. Fixed hourly pricing. No read rate limits. Always-hot data on de
`deepagents deploy` now supports user scoped memory! add a user/ directory in your project so each user gets their own writable AGENTS.md, seeded on first deploy and persisted across conversations. yo
deepagents now supports structured output for subagents! an under appreciated piece of context engineering is figuring out what context is passed back from subagents to the main agent. now you can def
Devin is now available in Windsurf. Local agents make you faster. Cloud agents do the work while you're not there. Windsurf 2.0 brings them together. Introducing Windsurf 2.0. Manage all your agents f
Devin runs in its own VM in the cloud. You plan locally, delegate to Devin with one click, and keep coding (or close your laptop). Included with every Windsurf plan. Access to Devin in Windsurf is rol
Document OCR benchmarks are still an open problem Existing document OCR benchmarks are either too narrowly focused on a specific type (e.g. FinTabNet, ChartQA), or on documents that aren’t reflective
Windsurf has released a new version of its AI-powered coding environment featuring the Agent Command Center, Spaces, and Devin integration. The update is available for download at windsurf.com/downloa
Engineering Is Magic is a post shared on X (formerly Twitter) referencing Elon Musk's well-documented enthusiasm for engineering and technological innovation. The content likely celebrates a remarkabl
ERNIE-Image is now in ComfyUI An open-source 8B DiT text-to-image model from @ErnieforDevs, licensed under Apache-2.0. Key highlights: - Open-source under Apache-2.0 license - Precise multilingual tex
Cristobal Valenzuela, co-founder of Runway, reflects on a recurring pattern in human thinking where each generation believes its current moment is uniquely transformative and historically definitive.
ComfyUI demonstrates practical examples of text rendering and design layout capabilities, showcasing how the platform can generate and manipulate typography within AI-generated images. The post likely
ComfyUI demonstrates multi-panel composition techniques in this second example post, showcasing how to generate images containing multiple distinct panels or sections within a single output. The examp
Final call. The Agent 4 Content Challenge ends tomorrow at midnight. 5K on the line. Make a video. Post it. This is your last chance. 20,000 up for grabs. 4 weeks. 4 winners. We’re launching the Agent
finally: @simonlast + @sarahmsachs on Latent Space! Notion has rebuilt Notion AI 5 times. This is the first time Simon has told the entire story. I've been trying to do this interview for ~3 years. We
Tesla's Full Self-Driving (FSD) version 14.3.1 demonstrated improved parking capabilities, successfully identifying the first available parking spot at a Costco location. The system efficiently select
FSD v14.3.1 review after 10 drives and many hours, here are my thoughts (it’s a great one) - v14.3.1 feels like a big jump even compared to FSD v14.3, even for “just” a point release build. Everything
Gemini 3.1 Flash TTS is rolling out in Google Vids and is available today in preview via the Gemini API and in @GoogleAIStudio. Whether you’re creating a pitch deck or recording a passion project, tra
Generative AI in nutshell: 'If the technology is as formidable as the [tech CEO's] claim, then they could be leading us toward existential disaster; if the technology proves less transformative, and t
Get Started in ComfyUI 1. Update ComfyUI to 0.19.1 (Desktop and Cloud will be available soon) 2. Go to Template → Search for ERNIE-Image Select the ERNIE-Image workflow 3. Download the missing models
Given the increasingly closed-source nature of the U.S. AI ecosystem, it is now more important than ever to push for the proliferation of open model and dataset releases. Datamule (@johngfriedman), @T
Good question This group calls for genocide against whites, supports the Islamic regime in Iran, and spreads antisemitism. Then they blame @elonmusk, yet never answer a simple question: why is South A
Grok, an AI assistant developed by xAI, recorded over 326 million web visits in a recent month, marking a new all-time high in traffic. The milestone suggests growing user adoption and interest in the
Grok just hit its highest monthly traffic EVER: over 326 MILLION visits in March alone That’s a massive 61% jump YoY and up 9.3% just since February People love Grok because it's the only AI they can
Have found the same things! Using glm-5 as a daily driver for a lot of things We've tested new OSS models the moment they're released for a while at Lindy. Inference is our #1 cost by a lot (more than
Have to make up for lost time WHAT FEATURES DID X AND XAI ADD IN 2025? 2025 is almost over and both X and xAI has have been cooking! Let’s see what features they worked on this year! Good job done by
Nous Research's Hermes is a series of fine-tuned large language models known for strong instruction-following, reasoning, and conversational capabilities. The post appears to be enthusiastic user feed
'Hey Grok' is a voice command feature that allows Tesla users to ask questions or set location-based reminders by saying the wake phrase. The feature can be enabled through the vehicle's Controls menu
Hey Republicans; He pardoned 1,600 violent criminals. You said nothing. He bulldozed the East Wing. You’ve said nothing. He’s interfered with the release of the Epstein files. You’ve said nothing. He
Researchers are making progress in decoding the communication systems of whales, with recent findings suggesting their vocalizations contain structural elements that function similarly to vowels in hu
Harrison Chase, co-founder of LangChain, was tagged or mentioned in a post by user @colesmcintosh expressing enthusiasm or confidence about deep agents ('deepagents 4 sure'), likely referencing advanc
I chatted with @ysmulki about MatX, chip design and where silicon designed for LLMs is headed (8:17) Tightly coupling SRAM and HBM on one chip (14:03) More MoE FLOPS, smaller KV cache load (16:08) Num
A post by Swyx on X (Twitter) humorously acknowledges that a minimal-effort post performed surprisingly well in terms of engagement, attributing its success to 'Maps Theory of Twitter' coined by @jose
I find it fascinating that OpenAI's Global Affairs team will denounce AI doomers for being too negative and then, simultaneously, advocate for an Illinois bill that would shield them from liability if
I have always said that the Cybertruck has a smooth ride because of the suspension. Now I prefer your description. “It feels like you are floating.” 🤗🤗 It is not spoken about enough how comfortable it
I joined @cognition to build this. Cloud agents are going to be doing most programming but developers still deserve powerful tools to stay in flow while managing them. Devin + Windsurf brings the best
I ran the same prompt for a London Estuary accent, a Newcastle accent and an Exeter, Devon accent - all three audio files are now embedded in my blog post https://simonwillison.net/2026/Apr/15/gemini-
I spent some time trying to distill all the complex factors impacting open models -- economics, capabilities, distribution, policy, etc. -- into a clear list of beliefs. Here they are in full. 1. It’s
📽️ IC LoRAs: Control Your Video 🎤 Hosts: Purz & Julien ⏲️ April 15th – 3pm PDT / 6pm EST 📍 Live on YouTube, X, and Twitch LTX 2.3 + IC LoRAs = video control unlocked. Outpainting, character stylizatio
IC LoRAs are a technique for controlling video generation, likely enabling fine-grained stylistic or motion control through specialized LoRA (Low-Rank Adaptation) models applied to video workflows. Th