Baseten now rents out the full RL training loop, NVIDIA is speeding up the weight refresh that keeps training in sync, and a production RL recipe plus a preprint runtime are tackling the two hard parts: learning from live traffic and running many agents at once.
Four preprints and two field reports point the same way: retrieval is moving from whole-document search toward personal memory, token-level evidence and governed context that outlives a single prompt.
A 22 MB vision model, a 0.8B model taught a new skill by rewriting one weight matrix, and quantization that keeps larger models close to full accuracy point to capable AI on small hardware.
New benchmarks show AI agents getting measurably better at routine wet-lab work, while the people they are meant to help still get bad advice and the safety checks stay uneven.
Patching cannot keep up with machine-speed exploits, so control is shifting to the moment before an agent acts: allow, constrain, approve, block or halt, with privacy and resilience built into the same layer.
A leveraged sell-off, permitting uncertainty for data centers, and warnings of unsustainable valuations have flipped the AI trade from chasing the rally to avoiding the wreckage.
A 2M-image dataset with 100 forgery methods, detectors that fail when the background is restored, and identity generation split from synthesis point to attribution instead of a yes-or-no verdict.
Preprints show agents generate executable mocks that misbehave, fail most long-horizon tasks, and need deterministic tooling, while security teams already know a closed ticket is not a fixed bug.
Tavus is replacing scripted phone menus with video agents that read tone and timing, while Lorikeet's research shows most customers who leave heard nothing first. The lesson is not just a better voice: it is a system that notices, remembers and hands over to a human.
Four US model families passed 90% on a C++-to-Rust benchmark, efficiency methods cut training and serving costs, and Huawei's biggest campus still holds under a tenth of the largest US facility's capacity.
Open-weight models beat frontier models on a budget, an open-source pentest tool jumps from 2 findings to 17 with an expert plan, and preprints show red-teaming and task-tampering are now automated.
Vendors are answering trust concerns with certified agents, restricted autonomy, escalation rules, and full-population audit, as Gartner predicts over 40% of agentic AI projects will be cancelled by 2027.
Profound, a platform that helps brands show up in AI answers, says it raised 180 million dollars at a 1.8 billion dollar valuation and was named a Major Player by IDC. Rivals Pixis and AirOps offer different answers to the same question: how does a brand get cited when an agent does the searching?
Anthropic, GC AI, Tavus and OpenEvidence show the same pattern: general assistants are entering finance, law, insurance and medicine, but only inside clear limits on advice, action and who decides.
An agent runtime in production, a serving simulator, faster FP8 kernels and a candid measure of lost speedup: AMD's software for its own GPUs is starting to look like a stack developers can actually use.
Hundreds of small agents that evolve without decoding text, sibling harnesses that rescue failed rollouts, and a frozen predictor that learns from deployment evidence show agents getting better without bigger models.
Model routers that weigh eleven factors, config files that differ by editor, a 4,500-deep submission queue for Meta's agent, and privilege without credentials: the agent stack is acquiring its infrastructure layer.
New pangenome and long-read tools run far faster than the methods they replace, and one of them overturns the textbook picture of a tumour's epigenome.
A repressor that holds epithelial cells in place, a map of the heart's shared controls, a stem-cell state that appears on demand, and a noncoding RNA that times a gene's reply.
Kazakhstan put $100 million into crypto funds, Russia opened Bitcoin to ordinary savers with a cap, and a US firm kept buying, while Washington's new overseers signal tougher scrutiny.
Delek plans to repay an $850 million loan, Calumet retired $265 million of debt, and Plains, Ring and Global Partners all paired earnings with deleveraging or distributions.
ING calls the selloff an idiosyncratic repricing, Scotiabank tracks broad dollar strength into payrolls, and the Canadian dollar slipped 0.9%, while France's 2027 election hangs over the euro.
Roku completed its sale to FOX and delisted, Utz nears a take-private close, Marygold agreed a $2.00-a-share sale, and Sylvamo put its European business up for sale.
Moderna, Atmus, Bloomin' Brands, Sunstone and Elanco all lifted full-year guidance, pairing the raises with buybacks, dividends and, in one case, a quarterly loss.
A tech-led selloff sent investors into Treasuries in the week of November 2 to 6, even as strong payrolls pushed the chance of a December rate rise from 60% to 80%.
The blue sweep was the spark, the podcasts agree. Was it the trigger, the debt, or the spending itself? This week's shows argue over what actually broke the AI trade.
Untreated pancreatic cancer leans toward arrest over death, a stomach marker persists from premalignancy to cancer, and two drug targets work only when the cell's state is taken into account.
A blood marker for amyloid works better with locally tuned cutoffs, a platform trial borrows registry data to test more drugs, and one brain signal shows up across three kinds of psychiatric risk.
Deshaun Watson reported symptoms on Tuesday and could not practise fully all week. Cleveland's rookie gets the Week 9 start against the Saints with two completed passes to his name.
Washington switched on a fuel tax credit, Missouri funded rural clinics, Ohio paid for pipes, and California counted a decade of cleanup. State government spent the week on things you can touch.
Ice-sheet experts put the tail risk above earlier estimates, satellites show ammonia rising faster than the climate alone explains, and two studies map what flood and collapse warnings can and cannot do.