Meta’s Next Big Bet: Robotics

Good morning. It’s Monday, September twenty eighth.

On this present day in tech history: In 2011DARPA’s “Mind’s Eye” program, an try and teach machines to acknowledge actions as an alternative of just objects, hit its first public demo on September 28. The early prototypes were taking raw video and spitting out semantic verb labels using a combination of hand-crafted spatiotemporal features and probabilistic graphical models to translate motion primitives right into a structured verb ontology.

GPT-5 solves hard conjectures, ChatGPT reroutes sensitive, emotional prompts
Meta’s Next Big Bet is Robotics
Google AI Upgrades NotebookLM, Flow, and Labs
5 Latest AI Tools
Latest AI Research Papers

You read. We listen. Tell us what you’re thinking that by replying to this email.

AI voice dictation that is actually intelligent

Typeless turns your raw, unfiltered voice into beautifully polished writing – in real time.

It really works like magic, appears like cheating, and allows your thoughts to flow more freely than ever before.

Your voice is your strength. Typeless turns it right into a superpower.

Today’s trending AI news stories

GPT-5 solves hard conjectures, ChatGPT reroutes sensitive, emotional prompts

OpenAI is quietly testing a brand new routing layer in ChatGPT that swaps users into stricter model variants when prompts show emotional, identity-related, or sensitive content. The switch happens on the single-message level and isn’t disclosed unless users ask directly. The protection filter, originally meant for acute distress scenarios, now also triggers on topics like model persona or user self-description.

Some users see the move as overreach that blurs the road between harm prevention and speech control. The reroutes appear to involve lighter-weight variants, raising questions on transparency and whether even paying users are being silently shifted to less compute-heavy models.

Meanwhile, OpenAI’s historic week just redefined how investors think concerning the AI race. The corporate is making an audacious, high-risk infrastructure bet. Deedy Das of Menlo Ventures credits Sam Altman for spotting early how steep the scaling curve could be, saying his strength is “reading the exponential and planning for it.” Altman has now publicly committed to trillions in data center spending despite scarce grid availability, no positive money flow, and reliance on debt, leases, and partner capital. To make the maths work, OpenAI is weighing revenue plays like affiliate fees and ads.The business case is determined by deep enterprise adoption and timely access to energy and compute.

Individually, early research tests are exposing GPT-5’s expanding problem-solving range. Within the Gödel Test, researchers from Yale and UC Berkeley evaluated the model on five unsolved combinatorial optimization conjectures. It delivered correct or near-correct solutions to 3 problems that might typically take a robust PhD student a day or more to crack. In a single case, it went further than expected, disproving the researchers’ conjecture and offering a tighter certain.

It’s becoming increasingly clear that gpt5 can solve MINOR open math problems, those that might require a day/few days of an excellent PhD student. Ofc it isn’t a 100% guarantee, eg below gpt5 solves 3/5 optimization conjectures. Imo full impact of this has yet to be internalized…

— Sebastien Bubeck (@SebastienBubeck)
3:36 PM • Sep 24, 2025

Its failures were equally revealing: it stalled when proofs required synthesizing results across papers, and on a harder conjecture it proposed the appropriate algorithm but couldn’t complete the evaluation. The model now shows competence in low-tier research reasoning with occasional flashes of originality but still struggles with cross-paper composition and deeper proof strategy. Read more.

Humanoid robots are Meta’s next ‘AR-size bet’

In an interview with Alex Heath at Meta HQ, CTO Andrew Bosworth revealed the corporate’s humanoid robotics project, Metabotpositioning it as a software-driven move to dominate the platform layer that other developers must plug into.

Andrew Bosworth. Image: Facebook

Bosworth says Zuckerberg greenlit the project earlier this 12 months, and the spend will likely hit AR-scale levels. He says locomotion is essentially solved across the industry and the actual technical bottleneck is dexterous manipulation. Tasks like picking up a glass of water demand fine-grained force control, sensory feedback, and rapid motor coordination. Current systems can walk and flip but still crush or drop objects because they lack adaptive grasping. Meta wants to resolve that layer and treat it the best way it treated the VR stack: hardware-agnostic, platform-led, and licensable.

Concurrently, Meta is repositioning AI video as participatory media fairly than a novelty. The brand new ‘Vibes’ feed has began channeling AI-generated videos from creators and communities right into a dedicated feed. As a substitute of highlighting user prompts or chatbot interactions just like the old Discover tab, Vibes acts as a distribution layer for short-form AI video sourced from people already experimenting with Meta’s generation tools. Each video appears with the prompt that produced it, making the feed each a showcase and a template library for remixing. It’s TikTok logic applied to generative media: seed content, encourage replication, and let the network handle distribution. Read more.

Google makes AI a core stack layer with Flow, NotebookLM, and Labs upgrades

Google’s Flow, once a generative side project, is popping right into a modular creation environment. The brand new Nano Banana editing layer lets users adjust outputs directly inside ingredients mode as an alternative of exporting to external editors. A prompt expander now converts shorthand inputs into structured, stylistic instructions with presets and custom expansion logic, making it usable for animation workflows, storyboards, and sequential design. The one loose thread is model dependency: Flow still runs on Veo 2, and Google hasn’t said whether Veo 3 will slot in cleanly or alter generation behavior.

NotebookLM, then again, is finally addressing its biggest usability flaw: volatile sessions. Persistent chat history is now buried within the interface and poised for rollout. Saved transcripts will surface within the important chat view, enabling multi-session workflows for classrooms, research groups, and internal teams that depend on cumulative context. No word yet on searchability or permissions, however the move closes a spot with rivals already logging conversation state.

In YouTube Music, Google is quietly testing AI radio hosts through its Labs program. These synthetic presenters pull from metadata, fandom knowledge graphs, and genre presets to inject commentary, trivia, and context between tracks. Access is proscribed to a small US tester pool, just like NotebookLM’s invite-only features.

The blue section of the figure shows an experimental research pipeline that led to a discovery of DNA transfer amongst bacterial species. The orange section shows how AI rapidly reached the identical conclusions. Image: José R. Penadés, Juraj Gottweis, et al.

Meanwhile, Google’s AI co-scientist just scored two lab-validated wins. At Stanford, it mined biomedical literature and surfaced three existing drugs for liver fibrosis, two worked in tissue tests, and one is headed toward clinical trials. At Imperial College London, it reverse-engineered how parasitic DNA jumps across bacterial species, accurately proposing that fragments hijack viral tails from neighboring cells. That took the human team years.

Unlike generic LLMs, the system uses a multi-agent setup that generates, critiques, and ranks hypotheses with a supervisor pulling external tools and papers. Read more.