Congratulations on the launch. We have added @Machenta_ to the database of people and organizations shaping AI. turingtree.com/org/machenta

Good luck in your future endeavors @nathancgy4!

New org added to the database: Palaestra Research turingtree.com/org/palaestra…

Abdulrahman Alfozan was just fired from OpenAI for allegedly installing a hidden audio recording device in an executive conference room at the company. The Saudi engineer previously worked at Meta and co-developed Gradio, an open-source tool for generating web interfaces for AI models.

Who has actually shaped AI? We ranked the top 50 most-cited researchers. How many did you know? Full details: turingtree.com/top-50

Matt Clifford, who has served two different UK prime ministers on AI matters, has joined Anthropic in a new role of Managing Director, International Affairs. As part of Chief Global Affairs Officer Tino Cuellar's team, Clifford will lead the company's work with governments outside North America ("Europe, Asia-Pacific, and beyond"). In a LinkedIn post, he writes that he thinks it's clear that the decisions that will shape AI's impact should be made by industry, governments, and civil society working together. "People everywhere want real agency over how this technology is built and used in their societies, and they should have it. Anthropic takes that seriously, which is why I'm joining." In his work for 10 Downing Street, he has helped write the UK's national AI strategy and also served as an advisor. Before focusing on AI from 2023, Clifford co-founded Entrepreneurs First, an international talent investor and startup accelerator program, funded by, among others, Demis Hassabis, Nat Friedman, and John and Patrick Collison.

NEW MOVE: @tamaybes has joined Google DeepMind from Mechanize (Source: LinkedIn)

NEW MOVE: @thenanyu is joining OpenAI

Former OpenAI researcher Adam Tauman Kalai (@adamfungi) has co-founded a new AI safety nonprofit called Institute for Responsible Superintelligence (RESI) (@RESI_org). RESI is aiming to build the "scientific foundations" needed to make superintelligence safe by design. There are two other co-founders of the new research institute based in Cambridge, MA: - Shafi Goldwasser (@ShafiGoldwasser), a Turing Award-winning cryptographer, co-inventor of zero-knowledge proofs - Vinod Vaikuntanathan (@Vinod_MIT), cryptographer, professor at MIT, and Gödel prize winner There is also a research team of 16 people, including Daniela Rus, director of MIT CSAIL (@MIT_CSAIL). RESI differs from other AI safety initiatives by focusing on formal and design-oriented foundations such as precise safety objectives and architectures with analyzable guarantees.

Full video: youtube.com/watch?v=aMB8JWge…

Llion Jones, co-creator of the transformer architecture, has decided to drastically reduce the time he spends on it. (From his TED Talk "How competition is stifling AI breakthroughs", recorded in San Francisco on October 21, 2025) @YesThisIsLion @SakanaAILabs

Former Alibaba Qwen lead launches new billion-dollar lab. Here are the details you need to know. 👇 Five months after leaving the Qwen team at Alibaba, Junyang Lin has announced his new venture, Pragmatik Labs. The Shanghai-based company is focusing on, according to its website: 1) Digital agents that transform knowledge work 2) Physical agents that bring intelligence into real environments Pragmatik Labs has reportedly attracted investments of $100 million each from Gaorong Ventures and HCG (formerly Sequoia China). Joining the round is also Tencent with $20 million and the Shanghai Engine Fund with an undisclosed amount. The Information reported that the valuation is $2 billion. @JustinLin610 @pragmatik_labs For more updates on the AI ecosystem, give us a follow.

John Yang, former head of multimodal foundation models at ByteDance, has joined Adobe as its new Head of Research and AI Technology. "I'm excited to work with an exceptional team to help shape the next generation of AI that amplifies human creativity and unlocks new possibilities for everyone." The latest update to Adobe's image generation model Firefly Image 5 was in October last year. At ByteDance, Yang worked on both image model SeeDream 4.0 and video model Seedance 1.0. Maybe we'll see more frequent updates on Adobe's own models now? For more updates on the AI ecosystem, give us a follow.

What will John Jumper, the Nobel Prize winner, do at Anthropic? Taking a look at some of the previous comments for the American researcher, we might get a hint. "Using AI to increase the pace of scientific progress is a core part of Anthropic's mission," the Claude maker writes on its website. Following that, the hiring of John Jumper makes a lot of sense. The 41-year-old American was trained in physics and mathematics and later earned a PhD in theoretical chemistry. In a presentation for Y Combinator last year, he introduced himself like this: "I'm someone who has done some work in AI for science, who really believes that we can use the AI systems, these technologies and ideas to change the world in a very specific way: to make science go faster and enable new discoveries." After his initial studies in physics, he ended up working at a company doing computational biology: "How we get computers to say something smart about biology," as he puts it. Jumper loved it. It was fun and would also let him do what he thought he was good at, including thinking hard thoughts about the nature of the world and using it toward making medicines or enabling others to do so. After finishing his PhD at the University of Chicago, Jumper joined Google DeepMind to lead the AlphaFold team. The goal was to use technology to push the scientific frontier in an industrial setting with smart people and great computer resources. "With all that, you darn well better make some progress," Jumper said. "I think the guiding principle for me has been that when we do this work, ultimately we are building tools that will enable scientists to make discoveries," he said. Jumper recognizes ideas as a core component of scientific progress, and it's not necessarily about coming up with one key innovation. Many mid-scale ideas can add up to a transformative system. Jumper has also been a proponent of open source and science. AlphaFold 2 was open sourced in 2021. "When you build the right tool, when you solve the right problem, it matters and it changes the lives of people who are doing things that you would not do but building on top of your work." He knew his work really mattered when a special issue of Science came out, and three out of the four papers featured made extensive use of AlphaFold. "I think I counted over a hundred mentions of the word AlphaFold in Science and we had nothing to do with it. It was just people doing new science on top of the tools we had built and that is the greatest feeling in the world." He estimates that AlphaFold has made the field of structural biology 5% or 10% faster in total. "The amount to which that matters for the world is enormous and we will have more of these discoveries." Where AlphaFold was a specialized system with a relatively narrow scope, he thinks we're moving toward more general systems. He thinks the most exciting question in AI for science is how general it will be. "Will we find a couple of narrow places where we have transformative impact or will we have very, very broad systems? I expect it will ultimately be the latter." Maybe he will build that at Anthropic. The company is hosting a science event on June 30, which could offer an early indication of Jumper will be working on.

OpenAI just hired one of the most influential AI researchers of the modern era. After a two-year stint at rival Google DeepMind, Noam Shazeer is joining the ChatGPT maker as its new lead for architecture research. Shazeer is a co-author of the seminal Transformers paper "Attention Is All You Need" (2017). He also co-authored a groundbreaking innovation in the Mixture-of-Experts architecture that made it work at massive scale and is used broadly in the industry today. In 2021, he co-founded Character AI, but returned to Google three years later, where he had started his career, in a $2.7B deal. At Google, he served as the technical lead for Gemini alongside Jeff Dean and Oriol Vinyals.

Sarvam (@SarvamAI) just raised the largest venture round for an Indian AI company this year. It's a $300M Series B round at a $1.5B valuation led by HCLTech, a global IT services and consulting company. Joining the cap table is also Bessemer Venture Partners (@BessemerVP), alongside existing investors Peak XV Partners (@peakxvpartners) and Khosla Ventures (@khoslaventures) Based in Bengaluru, Sarvam develops foundation models across a range of modalities: text-to-speech, speech-to-text, language, and vision. A key product is conversational AI agents built for India, supporting 11 languages in the world's most populous country. The funding will support Sarvam's work on a new frontier model for agentic AI, coding, and cybersecurity use cases, according to the press release. Sarvam AI was founded in August 2023 by Dr. Vivek Raghavan (@vivek_raghavan) and Dr. Pratyush Kumar (@pratykumar).