The U.S. and China just agreed on a channel for AI incidents.
China's foreign ministry says the two countries will establish a bilateral incident channel and hold another AI dialogue in November. The White House calls it a "super intelligence" dialogue. Neither side has said which incidents would trigger the channel.
Even rivals in the AI race now see a need for a line they can use when something goes wrong.
Sources:
1. https://t.co/d7H77bHW68
2.https://t.co/6UXZXKO9UQ
Bots outnumbered humans on the web in May, says Cloudflare's CEO.
In a new Decoder interview, @eastdakota says @Cloudflare saw automated traffic pass human traffic earlier than it had forecast. His pitch: let publishers charge AI crawlers tiny fees rather than subsidize their traffic.
If agents become the main audience for websites, the web needs a business model built for machines.
Source: https://t.co/75uSygxer7
A $1.25B AI power plan just fell apart.
Crusoe won't be Boom Supersonic's launch customer for 29 gas turbines. Boom CEO @bscholl says turbines are no longer in Crusoe's near-term primary power mix; Crusoe confirmed the split.
AI infrastructure isn't constrained by chips alone. The power plan has to survive contact with the site.
Microsoft's AI agent gets a name, a role and an office identity.
@Microsoft says Copilot Autopilot can watch channels, follow up and keep working while you sleep. Code lets staff build apps from prompts. Both are previews, not general releases.
The enterprise AI race is moving from chat boxes to coworkers with permissions.
We’re building Copilot as a new OS for work that spans every model, every form factor, and every task. Today, we’re announcing our biggest update to Copilot to date, bringing four things together:
· Autopilot: proactive and long-running agent built for the enterprise
· Code: build apps with Copilot, hosted inside your company’s tenant
· Home: Chat + Cowork together
· Office: now fully embedded in Copilot (and Copilot embedded in Office, of course!)
Plus, you can invoke Copilot in Teams, and we’re introducing Today, a proactive experience that surfaces the most important information from across M365 without needing to ask for it.
The way we work is changing and so are our workflows. This update brings AI into that flow, from answering a question, to building an app, to getting work done on your behalf.
Google's next AI chip test is in orbit.
@Google says its Project Suncatcher prototype rides a SpaceX launch next week to test whether TPUs survive launch vibration, radiation and the thermal extremes of low Earth orbit. Parts can feel 50-100g on launch. The two-satellite laser-link test is slated for 2027.
This is a hardware survival experiment, not an orbital data center. But AI's power bottleneck is now pushing compute beyond Earth.
Source: https://t.co/EIDJbmEs9E
Watch today: AI CEOs at the UN Security Council.
Sam Altman and Dario Amodei are expected to address the Security Council on AI, per a UN spokesperson. Hugging Face's Clément Delangue and Yoshua Bengio also expected. Bloomberg says the meeting is Wednesday.
One day after Trump told the same building he won't "stifle growth of something that will be bigger than the industrial revolution."
The labs are asking the world for rules their own government just turned down.
@OpenAI@AnthropicAI
Source: https://t.co/99hteyMMvu
Apple's new enterprise pitch: skip the token bill.
New Mac Minis and Mac Studios shipped Tuesday, and Apple's pitch to companies is that they're cheaper than renting data centers. At launch they strung four Mac Studios together to run a trillion-parameter model - off one wall outlet.
The catch: IDC puts Apple at 4.6% of the enterprise desktop. Windows has 91.3%.
Every other AI company is selling tokens. Apple is selling a way to stop buying them.
@Apple
Source: https://t.co/eduMacbwCo
Your small business just got a 24/7 employee — one that costs almost nothing while it sleeps.
DigitalOcean's Managed Agents is now in public preview. The non-developer version: you don't build infrastructure, you start from a prompt in their blog. The agent runs in the cloud around the clock, auto-pauses when there's nothing to do, and resumes in 305ms exactly where it left off. Per-second billing means an hour of mostly-idle work costs $0.060 instead of $0.126. @digitalocean
The proof it's built for normal teams, not just engineers: @qencode put together a support-triage agent in one day. It reads each incoming request, weighs the urgency, and routes it — saving them up to a full day per week.
Pause it Friday, pick it up Monday with everything intact. One cloud, one bill.
The takeaway: agents stop being a developer toy the moment running one 24/7 costs less than a coffee a month and you don't need an engineering team to babysit it. That line just got crossed.
https://t.co/KckGb9F0PL
@MatthewBerman@bot@spacexai A bot to find fun things for my daughter for the weekend, daily digest of weather with dressing tips for the day and school lunch menu
@NebulaAI 1. Possible to add WhatsApp integration?
2. Integrations -> browse app’s, seeing a bunch of icons for services is not so user friendly… take additional navigation to figure out what the service is. A list view with a 1 liner of the integration service would’ve been much better
Introducing the first vibe coding platform that builds agentic AI apps.
Build AI that thinks, searches the web, executes code, and completes tasks end-to-end.
One-shots Cursor, Perplexity, Gamma, and more.
In celebration of our launch, we're giving away a free month to everyone who replies.
Repost and reply “Blink” below and we'll DM you the credits 👇