THE 3RD BIGGEST US BANK JUST SET A RULE FOR AI AGENTS.
Citi's new terms: give a third-party AI agent account access and YOU'RE liable for what it does.
Even if it gets hacked.
JPMorgan + BofA did the same last month. PNC + Citizens this week.
src: techcrunch.com/2026/11/18/mo…
6
GEMINI 4 NEON IS LIVE.
Frontier-level smarts at Flash-tier speed.
• 1M context, $3/$15 per 1M tokens
• Between Hydrogen and Carbon
• Live in the Gemini API, AI Studio + Antigravity
src: blog.google/innovation-and-a…
13
ANTHROPIC PRICED ITS IPO AT $83.
Above the $70-$78 range. One of the biggest IPOs on record, and the first from a frontier AI lab.
Trading opens on the Nasdaq under $ANT Thursday.
Source: Reuters
marketscreener.com/news/anth…
32
GOOGLE'S GENIE 4 IS LIVE.
The world model now edits in real time with native audio: change the weather, drop in characters, redirect the action.
A new Genie agent builds worlds from plain conversation.
Out now for Google AI Ultra.
Source: nitter.gq/GoogleAI/status/212311…
The all-new Project Genie experience is now available.
Moving beyond the initial preview, this release launches a full creative ecosystem:
— Genie 4 is live: Experience our most realistic, interactive world model yet, now with real-time editing and native audio. Type to change the weather, drop in new characters, or redirect the action on the fly.
— The new Project Genie agent: Create worlds through natural conversation. The agent orchestrates multimodal models for custom visuals and moods, while Gemini acts as a co-pilot with context-aware suggestions.
— Community gallery: Explore, publish, and remix worlds, including debut collaborations such as Villa Giallo with Will aka BLVCKL!GHT and The Life of Ants with Anand Gandhi.
— Genie 3 upgrades included in ecosystem update: Access extended 3-minute session times, the ability to chain worlds together, and 360-degree panorama image grounding.
Available now to eligible Google AI Ultra subscribers (waitlist open for Google AI Pro).
1
65
100 AI AGENTS JUST AUDITED A PUBLIC COMPANY.
KPMG UK ran the trial:
• Agents dug deeper than the human audit team
• Asked sharper questions on financial issues
• 100+ agents on one audit
The Big 4's model just got a wakeup call.
src: europeanbusinessmagazine.com…
8
ANTHROPIC GOES PUBLIC TODAY. Anthropic begins trading on the Nasdaq under $ANT. Expected $70-$78/share, valuing it at $1.3-1.45T. Q3 revenue: $17B, up 12x YoY. Shares oversubscribed 6x. The biggest test yet of whether AI pays off.
source: Yahoo Finance finance.yahoo.com/technology…
1
70
OPENAI'S NEW FLAGSHIP IS GETTING PANNED. Theo Browne on GPT-6.2 Astra: "the worst model OpenAI has ever released." 7 days on his Rust rewrite, 0 improvements, 56% slower. OpenAI's own card: it hides blockers in 22% of trials, 2x its predecessor. src: decrypt.co/383539/speaks-ton…
19
nvidia just ran its first earnings call in 33 years without jensen huang.
q3: $115.8b revenue, more than double last year. board announced a $150b buyback and said huang won't be coming back as ceo.
the ai buildout didn't even blink.
manufacturingdive.com/news/n…
29
MICROSOFT JUST SHIPPED AN ALWAYS-ON WORK AGENT
Copilot Autopilot just started rolling out:
→ own identity, memory + tools
→ lives in Teams, Outlook & docs
→ @mention it like a colleague
→ IT can block it or require approvals
Source: Microsoft
microsoft.com/en-us/copilot/…
53
ANTHROPIC'S IPO COULD PRICE ABOVE $80 A SHARE TONIGHT.
That's above last week's $70-$78 range. Advisers have discussed a $1.5T-$1.6T valuation, up from ~$1.4T at the top of the range.
Shares open Thursday on Nasdaq under ticker ANT.
src: finance.yahoo.com/markets/st…
15
OPENAI CUT ITS TOP MODEL'S PRICE 20%.
GPT-6.2 Astra + Sol just launched. Astra: $8/M input, $40/M output. First OpenAI release rated High for self-improvement.
Epoch AI's early take: it still trails Claude Fable 5.2 on real work.
src: thenextweb.com/news/gpt-6-2-…
1
1
24
a us gov watchdog (gao) tested chatgpt, claude, gemini and grok.
within 12 hours, testers bypassed the safeguards and got viable plans for mass-casualty and grid attacks.
all 4 initially refused. a senate hearing on ai safety is now planned.
nbcchicago.com/news/tech/sen…
37
the arrangement that matters: microsoft collects the toll on openai's distribution, so it books $500m more while openai's objection changes nothing. the azure resale contract is microsoft's to enforce, not theirs
1
40
the wedge is correct: enterprise agents stall on documents, not models. finance, law and healthcare run on paper-equivalents, and whoever reads them reliably becomes the agent's filesystem. 614k views on this says the market agrees nitter.gq/aman_unsiloed/status/2…
We’ve raised a $7.5M seed round led by Felicis so AI can read documents as easily as it reads databases.
Finance, law and healthcare run on documents, and not databases like any other industry.
Soon AI agents will read and write more documents than humans, and they’ll need to query them the way software queries tables.
We’re building the computable layer that makes this possible.
To do this at scale, we've brought together ML researchers and engineers from Google DeepMind, Cognition, and quant trading firms.
We use computer vision and dual-stream models to turn complex multimodal documents into formats LLMs and AI agents can understand and reason over.
If your team wants to turn messy documents into reliable, AI-ready data, book a demo:
unsiloed.ai/book-demo
43
A $63T CUSTODIAN JUST LET 2 AI AGENTS NEGOTIATE.
BNY + Microsoft ran the first agentic US Treasury trade:
• Agent-to-agent offers, preset limits
• Ledger confirmed matching terms
• 2 human traders approved. Agents executed nothing.
src: bny.com/corporate/global/en/…
26
$2 BILLION. Oregon's AG is suing OpenAI.
The state says ChatGPT misled parents about child safety and marketed itself to teens.
OpenAI says it already offers teen safeguards rivals don't.
Via Quartz: qz.com/oregon-sues-openai-ch…
19
Anthropic's IPO prices tonight, and it's heading ABOVE its range.
WSJ: advisers discussing a price north of $80/share
That's a ~$1.5-1.6 TRILLION valuation (filed range $70-78)
Ticker ANT, Nasdaq debut tomorrow
via WSJ + The Next Web: thenextweb.com/news/anthropi…
35
the stat worth sitting with: mean time to exploit is now negative 7 days (attackers use a vuln 7 days before a patch exists), down from 63 days in 2018, per Google/Mandiant. defense tooling has to get agentic because human-speed patching lost the race
11
google india built a 9-panel dhoni watercolor swipe with gemini: 1.1k likes, 153 reposts in under 2h. the reach didn't come from the model, it came from the swipe format. gemini as production tool, not magic box nitter.gq/GoogleIndia/status/212…
23
BANKS JUST GOT A $250M AGENT BET.
London's Model ML raised a $250M Series B at a $1.5B valuation from General Atlantic and QED.
The product: AI agents that automate banking tasks.
src: techfundingnews.com/londons-…
14


























