#Other

AIChip Huyen shared a casual post saying she was bored at the airport and built killedbygpt.com. The post provides no further details about the site's purpose or features.

AIMicrosoft CEO Satya Nadella joined swyx, Sarah Guo, and Elad Gil for a wide-ranging podcast episode covering AI, platforms, builders, and future directions. The episode was released as a special MSBuild crossover pod.

AIMaintainers and NVIDIA engineers improved multi-GPU performance in ggml, the low-level engine behind llama.cpp, yielding significant gains on RTX systems. The work also lays groundwork for hardware-agnostic tensor parallelism in ggml. Details are in a technical blog from NVIDIA RTX Spark.

AIAmazon CEO Andy Jassy joined Amazon teammates to prep meals at Community Lunch on Capitol Hill in Seattle during Global Month of Volunteering in May. The nonprofit serves free, made-from-scratch meals to 200-300 low-income and unhoused neighbors each weekday, and nearly 2,000 Amazonians have volunteered there since 2023. Amazonians completed more than 500,000 volunteer activities across 55 countries last year.

AIMax Woolf reports that Hy3, an unfamiliar LLM, has unexpectedly topped OpenRouter's AI Model Rankings by a large margin. After examining the usage data, he says it only deepened his confusion about the result. The full analysis is in his new blog post on minimaxir.com.

AIGoogle Labs has refreshed its homepage at labs.google so visitors can more easily find its newest innovations, including those announced at I/O. The site lets users explore the portfolio, try experiments, and learn more about Google Labs.

AIGoogle Labs is promoting a set of experimental features available at labs.google. The post offers no further details about the features' functions or availability.

AIGoogle Labs is promoting its latest AI experiments and inviting users to explore them at labs.google and help shape the future of AI technology. The post also asks users to share what they create with the experiments.

AICognition has released Auto-Triage in Devin Automations, which lets Devin respond to Slack messages, Linear events, GitHub activity, schedules, and webhooks. Devin can investigate with connected observability tools and the codebase, then post a summary, tag an owner, or open a PR. Devin runs in network-sandboxed environments with added protections against prompt injection and data exfiltration, and a limited-time offer gives $200 in credits for a first automation.

Why it matters: The post shows how an agent handles alerts and bug reports from existing team channels, a practical pattern for teams weighing automated incident response.

AIMira Murati argues that today's AI experience is turn-based, forcing people to batch their thoughts, avoid pointing at things, and phrase questions like emails. She says the interface leaves no room for users, so people end up adapting to the models rather than the reverse.

AIOpenAI's Fidji Simo praised CellularIntel as a strong example of AI applied to biology and said she is proud to advise the team. The post cites the company's progress from an early idea, and the referenced post reports CellularIntel acquiring the STEM-PD Phase 2-ready allogeneic cell therapy program for Parkinson's from Novo Nordisk, which will also invest in CI.

AINemotron Hackathon Winners - S Korea

AIChip Huyen congratulated Alec Radford, Luke Metz, and Soumith Chintala on winning an ICLR Test of Time award for work published at ICLR 2016. She highlighted that the team's paper was produced by researchers in their twenties without any PhDs among them, and expressed hope that the three collaborate again.

AINVIDIA AI Developer thanked @Picomoorhead for a contribution to OpenClaw, which the post links to via @openclaw. The post provides no details about what the contribution was or what it enables.

AIAwni Hannun invites attendees to Ollama's MLX meetup at Ollama's Palo Alto office this Thursday, April 9, at 6pm. RSVP is required because space is very limited, and food and drinks will be provided.

AIKarpathy describes using LLMs to compile raw source documents into a markdown wiki that he views in Obsidian, with the LLM writing and maintaining most of the wiki. He reports that at about 100 articles and 400K words, the LLM agent can answer complex questions directly from the wiki, and he also runs LLM health checks to find inconsistencies and gaps. He shares the underlying idea as an "idea file" that users can give to their own agents to build a customized version.

AIAndrej Karpathy describes using LLMs to compile raw research sources into a markdown wiki of about 100 articles and 400K words, viewed in Obsidian. He says an LLM agent answers complex questions against the wiki without RAG, with outputs filed back to enhance it. He also suggests the workflow could become a product rather than a collection of scripts.

AIAndrej Karpathy says memory-based personalization in LLMs can be distracting, as a single question from two months ago keeps resurfacing as a supposed deep interest. He describes this as the model trying too hard.

AITri Dao announced Mamba-3, which he described as the most powerful linear sequence model to date, as hybrid architectures increasingly rely on strong linear models. The post cites Qwen, Kimi-Linear, and NVIDIA's Nemotron-3 Super as examples of this trend. According to co-author Albert Gu, Mamba-3 shows noticeable performance gains over Mamba-2 and Gated DeltaNet at all sizes while maintaining speed.

AIChris Olah, the account owner of which is Anthropic, shared two Greenhouse job listings and invited applications. The post itself gives no details about the roles, teams, or requirements.

AIGoogle DeepMind researcher Oriol Vinyals announced he is moving back to California after ten years in London, to keep working on Gemini and toward AGI. The post is a personal update with no product, benchmark, or technical details.

AIMistral announces Voxtral 2 with two transcription models: Voxtral Realtime, released under an Apache 2 license with latency configurable to sub-200 ms, and Voxtral Mini Transcribe 2, which adds speaker diarization, word-level timestamps, and context biasing. The models support 13 languages and are available through the Mistral API, which the post describes as one of the most cost-effective transcription APIs on the market. The attached chart shows word error rates on FLEURS across Italian, Spanish, English, German, Portuguese, French, Russian, Dutch, and Chinese at several latency settings.

AIMistral's robotics team is recruiting, with the account owner inviting readers to join. The post gives no details on roles, locations, or requirements.

AILilian Weng says she enjoys working with people who care about craftsmanship and what they build. She describes having the chance to work on something she is passionate about, beyond earning a living, as a privilege she does not take for granted.

AIGeoffrey Hinton says he recently talked about the old days with Jeff Dean, in a fireside chat moderated by Jordan Jacobs. The full recording of that discussion is now available on Spotify.

AIKarpathy argues animal intelligence is only one point in a large space of possible minds, and LLMs arise from a fundamentally different optimization process. He contrasts survival-driven animal drives with LLM training shaped by imitation of human text, RL on task distributions, and user engagement metrics, which he says leaves LLMs jagged and prone to sycophancy. He calls LLMs humanity's first contact with non-animal intelligence and says people who build accurate internal models of them will reason about them better.

AIThinking Machines is recruiting, with research roles now live on its Greenhouse job board and more positions planned. The main post simply invites readers to consider joining the team via the company's careers link.

AIGoogle DeepMind researcher Oriol Vinyals is traveling to Barcelona to receive a Doctor Honoris Causa from his alma mater, UPC. He also announced a Thursday master class titled "From AI to AGI: The Quest for True Intelligence," with registration details linked in the post.

AIGeoffrey Hinton recorded a podcast with Jon Stewart, whom he describes as a longtime hero, focused on explaining how AI works. Hinton says Stewart was especially eager to understand the underlying mechanics.