three ways agents got access to money in november. chatgpt wallet pays with a network token after a face id or fingerprint check on every purchase. claude personal finance reads accounts through plaid and can't move a cent. poppy 1.0 scopes access per company. if a site can't, web.unscoped gives the agent full access.
2
45
an ai agent ran for 48 hours straight and came out with a chip coolant molecule that passed a real lab test. alibaba says qwen4-max did it in four closed-loop rounds: molecular design, synthesis and testing, active learning. 2.4t parameters, 71b active per token. cached input: $0.075 per million tokens.
1
1
66
gemini 4 hydrogen, google's fast tier, scores 51 on the artificial analysis index. argon, a frontier model, scores 50. in antigravity, orchestrator pairs it with carbon: carbon plans read-only, hydrogen edits and runs tests, and hands back to carbon when it gets stuck. google's number: near-carbon task success at 24% less quota.
1
34
google's antigravity post: antigravity.google/blog/gemi…
26
the microsoft 365 outage wasn't a traffic spike. servers kept failing automated health checks and dropped out, and the rest buckled under their load. restarted servers went unresponsive again as soon as traffic hit them. 21 hours in, openai's microsoft 365 connector is still listed as broken. agents that read mail through it are stuck too.
1
33
microsoft's status thread: teapawt.org/MSFT365Status/status/2… openai's status page: status.openai.com/history
Replying to @MSFT365Status
We determined the elevated load isn't driven by an increase in traffic, but by a number of servers being unexpectedly unavailable; straining the remaining infrastructure. We're working to safely return them to service. More info under EX1550543 and MO1550821 in the Microsoft 365 admin center, as well as status.cloud.microsoft if the admin center is not available to you.
2
each anthropic engineer now ships over 45x more code a day than in 2024, its ipo filing says. claude wrote over 90% of september's contributions to the codebase. headcount grew 5x, quarterly revenue 86x. enterprise users without claude code averaged under $75 a month; the top 5% with it, about 20x that.
1
36
1,000 claude fable 5.2 agents formalized fermat's last theorem in 28.7 hours. they started from 47 lemmas proved earlier, and opus 5.5 in the same setup was 26% done at hour 18. anthropic didn't publish the bill: about 2 billion output tokens, roughly $80k at list price. that's not a max-plan weekend.
1
19
math: the earlier run used about 6 billion output tokens, this one a third of that, at $40 per million output tokens. anthropic's write-up: claude.dev/blog/safely-scali… · launch post: anthropic.com/claude-fable-a…
6
a model right on 97% of answers still failed 37% of documents. artificial analysis's decision index v2 counts a document only when all 15 answers are right. drex max v1.1: 97% per answer, 63% per document. 0.97^15 is 63%. claude haiku 5.5: 99% and 92%. for an invoice agent, the document is the unit.
1
1
33
artificial analysis, decision index v2: artificialanalysis.ai/articl…
13
an unreleased anthropic model helped close an 18-year-old quantum conjecture. physical superintelligence's research system had a lean-accepted proof on day three, then improved it twice. the upshot: on random inputs, a quantum computer can't save exponentially many lookups unless the input hides a rare pattern.
1
38









