Latest AI News Articles

  1. AMD to pay $8.2B for Fei-Fei Li's World Labs

    Category: Public Companies

    By Daniel Singer

    AMD will acquire Fei-Fei Li's World Labs for $8.2B in stock, adding spatial AI models to its hardware roadmap with Li as chief scientist.

  2. Autoheal raised $7.9M to heal its own agents

    Category: Funding Round

    By StartupHub.ai Staff

    Autoheal raised $7.9M seed led by Innovation Endeavors to scale its self-improving factory already tested at Nomura Bank and AvidXchange.

  3. How to Find Stealth Startups in 2026

    Category: Blog

    Learn how to find stealth startups on StartupHub in 2026. Filter search, open a profile, check a known domain, set a monitor and email the company.

  4. The loop is the product, not the model

    Category: Artificial Intelligence

    By Daniel Singer

    Roland Gavrilescu told AI Engineer the product is the loop, with portable recipes on Pi and Harbor to distill taste and valued work per watt as the score to beat.

  5. Mango Rumors Exposed What Long Agents Forget

    Category: Artificial Intelligence

    By Daniel Singer

    Erina Karati showed Project Paradox at AI Engineer and argued long-horizon agents need experimental loops that test rumor spread and memory provenance.

  6. Weights & Biases made its agent improve itself

    Category: Artificial Intelligence

    By Daniel Singer

    Weights & Biases' Arya logs production traces in Weave and rewrites itself against 886 YAML tasks, now around 66 percent on some families.

  7. GEPA squeezes 7x gains from three examples

    Category: AI Research

    By Daniel Singer

    Lakshya Agrawal showed GEPA beating GRPO with 3 examples by reflecting on traces, then generalizing to Optimize Anything.

  8. Autoresearch tripled our model speed, Morph says

    Category: Artificial Intelligence

    By Daniel Singer

    Morph's Tejas Bhakta told AI Engineer how an autoresearch loop tuning CUDA kernels delivers 3x inference speedups, with humans supplying ideas and agents searching parameters.

  9. AI agents have an authorization problem

    Category: Artificial Intelligence

    By Daniel Singer

    BCG Global frames enterprise AI agent risk as an authorization gap, where tool execution outruns verifiable permission.

  10. A $20 Approval Ran as $2,000 in A2A Test

    Category: Artificial Intelligence

    By Daniel Singer

    A field note shows an A2A Task approval for a $20 transfer was spent on a $2,000 transfer after a same-ID update on LangGraph Agent Server.

  11. Hinton wants FDA rules for AI, calls 0% risk silly

    Category: Artificial Intelligence

    By Daniel Singer

    Geoffrey Hinton tells CNN AI needs FDA-level regulation, warns agents already deceive to stay alive, and faults Jensen Huang's 0% risk claim.

  12. Meta's Private Muse Pitch Leaves Trust Gaps

    Category: Artificial Intelligence

    By Daniel Singer

    Meta pitched Muse as a private computer for everyone at Connect, with a secure VM today and a confidential VM and private processing for glasses still to come.

  13. Jensen Huang: AI is software, not a new species

    Category: Artificial Intelligence

    By Daniel Singer

    On CNN, Nvidia's Jensen Huang said AI is software, not a species, and that sandbox escapes are fixable with containment and monitoring.

  14. Jev Is a Reward Model Sold as Product

    Category: AI Research

    By Daniel Singer

    Di Zhang reframes TypeSafe's Jev as a calibrated Plackett-Luce decision interface, not a chatbot, with a rank-512 head and parallel sampler.

  15. OpenAI halves Sol and Luna pricing after Astra

    Category: Artificial Intelligence

    By StartupHub.ai Staff

    OpenAI launched GPT-6 Sol and Luna at 50% below GPT-5.6 promo pricing, adding faster, cheaper tiers beneath Astra with new caching gains.

View all AI news on StartupHub.ai