0:08
Humanity’s Last Exam is a Distraction
This article takes a gentle dive into the ultimate AI systems evaluation benchmark, outlining why it was created, curating diverse opinions from groups of experts in the field about it, and wrapping up with a summary of the most widely accepted verdict.
0:24
OpenAI reportedly offers the Trump administration a five percent stake in the company
OpenAI’s board is apparently drafting a deal that would hand a five‑percent equity slice to the Trump administration, something you don’t see every day in a tech company’s playbook. The idea isn’t just a cash infusion; it’s a structural tie that could give Washington a formal seat at the table when OpenAI charts its research and product roadmaps.
What’s odd is that the terms are still fuzzy—there’s no public word on whether the government would contribute funding, policy support, or something else entirely. The move suggests OpenAI wants a more permanent channel to influence, or at least stay aligned with, federal AI strategy, rather than the usual lobbying or advisory roles.
If the stake goes through, it could reshape how the company navigates regulatory scrutiny, especially around safety standards and export controls. It also raises questions about how much policy direction might flow back into the model development pipeline.
All of this feels like a subtle, but significant, shift in the relationship between a leading AI lab and the political sphere—one that could set a precedent for how private AI firms embed themselves in national agendas.
1:25
Google brings TikTok-style video shorts to NotebookLM
Google has expanded NotebookLM to include video overviews in the short format commonly used on social media.
The article Google brings TikTok-style video shorts to NotebookLM appeared first on The Decoder.
1:39
AI agents can now complete 16 percent of freelance jobs at pro quality, up from 2.5 percent eight months ago
The Remote Labor Index measures how often AI agents complete paid freelance projects at professional quality. In eight months, the top automation rate has more than quadrupled.
The article AI agents can now complete 16 percent of freelance jobs at pro quality, up from 2.5 percent eight months ago appeared first on The Decoder.
2:02
Nvidia is bankrolling AI startups to loosen Big Tech's grip on its chip business
Nvidia is increasingly acting like a central bank for AI startups, actively shaping the compute market.
The article Nvidia is bankrolling AI startups to loosen Big Tech's grip on its chip business appeared first on The Decoder.
2:18
The Untaught Lessons of RAG Question Parsing: Structure Before You Search
Enterprise Document Intelligence [Vol.1 #6ter] - Six positions on the question-parsing brick that contradict the mainstream RAG playbook
The post The Untaught Lessons of RAG Question Parsing: Structure Before You Search appeared first on Towards Data Science.
2:36
2026 BAIR Graduate Showcase
Congratulations to the Berkeley Artificial Intelligence Research (BAIR) Lab class of 2026! This year, BAIR celebrates another remarkable group of Ph.D. graduates whose curiosity, creativity, and perseverance have pushed the frontiers of artificial intelligence and machine learning.
Their work spans the breadth of modern AI — robotics and embodied intelligence, large language models and reasoning, computer vision, generative modeling, AI safety, human-AI interaction, AI for science and healthcare, and much more. Along the way, they have published influential research, built systems with real-world impact, mentored their peers, and shaped the BAIR community for the better.
Now they are headed everywhere ideas travel: to faculty and postdoctoral positions, to industry research labs, and to startups of their own founding — and several are still exploring what comes next and would love to hear from you.
Please join us in celebrating the achievements of these wonderful graduates. We are proud of everything they have accomplished at Berkeley, and we can’t wait to see what they do next!
Thank you to our friends at the Stanford AI Lab for this idea!
Email: [email redacted]
Website: https://bfshi.github.io/
Advisor(s): Trevor Darrell
Research Blurb: I work on building generalist vision and robotic models.
What's next: Member of Technical Staff at Physical Intelligence
Email: [email redacted]
Website: https://sea-snell.github.io
Advisor(s): Dan Klein
Research Blurb: My work aims to understand when and how the different LLM scaling paradigms can be traded off and interchanged. In particular, test-time scaling treats each prompt independently, drawing long chains of inferences and then forgetting them entirely between prompts. This differs critically from pretraining, which instead learns a compressed representation from a large dataset. I believe bridging the gap between these methods of scaling computation, presents a key open challenge in the field: how can we develop methods which turn the inferences drawn at test-time back into learned representations that the model can hold onto across interactions.
Email: [email redacted]
Website: https://devinguillory.com
Advisor(s): Trevor Darrell
Research Blurb: Accounting for data shifts in computer vision models
What's next: Building collaborative AI systems, looking for conspirators.
Email: [email redacted]
Website: https://efleisig.com
Advisor(s): Dan Klein
Research Blurb: I design language models to work reliably and fairly for the broad range of real LLM users. First, my research leverages disagreement among user preferences as signal, in order to train and evaluate LLMs for entire populations of users. Second, I work on designing rigorous evaluations to extricate challenging LLM harms that diverse users face. Finally, I work on core technical failures of LLMs, like miscalibrated confidence, to reduce downstream risks when models are deployed to users with different needs. Combined, these interventions facilitate building LLMs that minimize societal harms, and maximize benefits to a wider range of real-world users.
What's next: Postdoctoral fellow at Princeton CITP
Email: [email redacted]
Website: https://graceluo.net
Advisor(s): Trevor Darrell
Research Blurb: My research is on interpreting and controlling generative models. For example, I've worked on re-purposing image generators for computer vision tasks, and meta-modeling language activations for better LLM probing and steering.
What's next: Research scientist in industry
Email: [email redacted]
Website: https://hanlinzhu.com/
Advisor(s): Stuart Russell, Jiantao Jiao
Research Blurb: My research centers on understanding and improving the reasoning capabilities of large language models (LLMs).
What's next: Member of Technical Staff at OpenAI
Email: [email redacted]
Website: https://haozhi.io/
Advisor(s): Jitendra Malik, Yi Ma
Research Blurb: Dexterous Manipulation and Robot Learning
What's next: Research scientist at Amazon; Faculty at University of Chicago
Email: [email redacted]
Website: https://zamfi.net
Advisor(s): Bjoern Hartmann
Research Blurb: My research focuses on effective human-AI co-design. I study the boundaries of language interfaces as a medium for interacting with AI, creating systems that blend language-focused interactions with structured user interfaces that draw on different levels of abstraction. I focus on language-oriented technologies, like LLMs and text-to-image models, that are powerful mediators of design processes.
6:12
Ashton Kutcher leaving Sound Ventures to launch new VC firm with Morgan Beller
Sound Ventures was always about putting big bets on AI labs that seemed poised to dominate their fields. Kutcher’s departure isn’t just a career move; it’s a shift in focus from the headline‑making models to the plumbing that keeps them running—things like compute hardware, data pipelines, and the energy systems that power the crunch.
With Morgan Beller joining forces, the new fund is positioning itself as a back‑stop for the infrastructure layer, betting that the real value will emerge from the tools and platforms that let AI scale. They’re looking at everything from chip design to low‑carbon data centers, aiming to tighten the supply chain that fuels the next wave of intelligent applications.
The idea is simple: if you can make the engines more efficient and reliable, the AI models built on top of them will get cheaper and faster, and the whole ecosystem benefits. It’s a quieter, more technical angle, but one that could reshape where the money flows in the coming years.
7:05
SpaceX shows investors a slim AI smartphone prototype powered by xAI technology
SpaceX showed investors a prototype AI smartphone that's supposedly thinner than an iPhone and integrates xAI tech. The device runs on a Qualcomm Snapdragon chip with its own operating system. Musk wants to build an "everything app" modeled after WeChat.
The article SpaceX shows investors a slim AI smartphone prototype powered by xAI technology appeared first on The Decoder.
