Morning Edition · Saturday, October 31, 2026
Tech
Anthropic Reports 17 Lab-Validated Protein Binders Designed by Claude, and Says It Is Not an Antivenom
An internal Claude research model generated 318,484 candidate designs over 64 hours using about 10,000 GPU hours, and 211 designs, along with experimental results, are being released.
Tech
Berkeley Lab Cut Beamline Image Analysis From About a Month to Roughly 15 Minutes Using Meta's Open Vision Models
The multi-lab SYNAPS-I project fine-tunes Segment Anything 3 and DINOv3 on 300 NVIDIA A100 GPUs inside government computing environments, keeping the data on federal infrastructure.
Tech
OpenAI Says It Made Over 200 Law Enforcement Referrals in the Third Quarter and Will Publish a Safety Report From Early 2027
ChatGPT will now automatically end conversations in which a user appears fixated on violence, and accounts may be paused after repeated incidents. A watchdog group disputes the effectiveness of the platform's teen guardrails.
Tech
Android Studio Opens the IDE to Third-Party Coding Agents Through an Open Protocol
The stable Rabbit 2 release lets Claude Agent, OpenAI Codex and Google Antigravity control the IDE over the Agent Client Protocol, passing the project graph, build diagnostics and emulator controls to the agent.
Tech
A Startup Says Measurements From Lab-Grown Neurons Made an Open Video Model Five Times Faster, Running on Standard GPUs
The Biological Computing Co.'s mischief 1 is built on the open-weight Wan 2.2 model and priced at half a cent per second of video, and no neurons are involved at inference time.
Tech
Bloomberg Identifies Anthropic as SpaceX's "Customer B," the Buyer Behind 34.5% of Its Third-Quarter Revenue
Anthropic's own IPO filing shows it can leave its compute contract with SpaceX, worth up to $120.5 billion, on 90 days' notice after set dates in January and March 2027.
Tech
OpenAI, Google, Microsoft and Anthropic Agree to Honor One Robots.txt Token for Opting Out of AI Training
Anthropic says it has already started implementing the shared signal. Microsoft is aiming for February 1, 2027, and Google says only that its support will come in 2027.
Tech
Former Anthropic Researchers' Startup Mirendil Releases Fern 0.1, a Coding Model With a Claimed 10-Million-Token Context
On the RULER long-context benchmark at 10 million tokens, Mirendil reports 100% on single-needle retrieval but 45% on multi-hop question answering.
Tech
Ant Group's Inclusion AI Reports 5.8x to 6.9x Decoding Speedups From a Parallel Draft-Tree Method Called UBTree
The drafter builds trees of up to 64 candidate tokens that the target model checks in a single forward pass, and it beat the DARTree baseline in all 28 reported comparisons.
Tech
Anthropic Renames Claude Max 20x as "Ultra" and Removes the Five-Hour Usage Window for Verified Subscribers
Prices stay at $100 and $200 a month and weekly caps remain, but new Ultra accounts keep the five-hour limit until identity verification clears.
Tech
Bank of America Tells Customers That Sharing Logins With AI Agents May Leave Them Liable for the Agents' Transactions
The bank's October 27 email warns that an agent's activity could trigger fraud locks on an account, and a spokesperson confirmed the message is authentic.
Tech
SETI Institute and Breakthrough Listen Let AI Agents Steer the Allen Telescope Array From Natural-Language Requests
The pipeline runs on NVIDIA's Holoscan sensor-streaming platform and has detected the Crab Nebula pulsar in real time. That pulsar is a well-known source, which makes the detection a validation of the system rather than a discovery.