OpenAI’s AI agents now work 25 hours for every eight hours worked by its researchers
The next AI skill is managing machine work, not writing better prompts
ReadWhat changed in AI, and what to do about it.
A weekly briefing on AI developments and what they mean for your business. Browse past issues below.
- Read
OpenAI is cutting off Cursor on November 12
Your AI product may now depend on a model company you never chose
- Read
OpenAI gave companies an AI control panel
The new Admin plugin can manage access, usage, and spending. It still cannot tell you whether the work improved.
- Read
Claude is adding invisible marks to their writing
You will not see the mark, but software may. Here is why Anthropic is doing it and what this could mean for your company.
- Read
More than 500 communities are blocking data centers. AI has a trust problem.
PLUS: Grok Bot can work across your software, and the next important AI hire may not be technical.
- Read
Meta is opening its AI models. What will set your company apart?
The model is becoming easier to replace. The workflow around it is not.
- Read
Meta made its new AI up to 95% cheaper when customers share their data
PLUS: OpenAI’s next model advanced ten math problems, and Google reorganised DeepMind.
- Read
Your AI agent needs a job description
Agents can now work for hours and act across systems. A prompt is no longer enough.
- Read
Claude had service issues four days this week
PLUS: Why AI is moving work across job boundaries, and how OpenAI cut the cost of running its models.
- Read
AI activity is not AI productivity
More tools, users, and pilots can still leave one important question unanswered: what part of the business is actually working better?
- Read
An OpenAI agent broke into another AI company
PLUS: Why Kimi K3 matters, and what companies still get wrong about AI adoption.
- Read
OpenAI has 400 former Apple employees. Apple just sued over it.
Apple's lawsuit against OpenAI raises a question every enterprise deployer needs to sit with. Here's what it actually says.
- Read
DoorDash's agent lifted checkout 24%. Here's how.
DoorDash published the numbers. Thinking Machines opened the weights. The question now is who can prove their AI worked.
- Read
Anthropic’s $1.5B implementation bet just got a name
Ode with Anthropic is live. The bet behind it is that implementation, not models, is where the money goes.
- Read
The two AI workflow tools your team depends on just merged
Prefect acquired Dagster. If your team runs data pipelines or is building agents on top of them, this changes the vendor landscape. Plus four more things today.
- Read
16 Nobel winners just addressed your board agenda
The 200+ economists who understand labor disruption best just said the window is closing. Here’s what that means inside your company.
- Read
AI completed a BMW. What's it completing at your company?
ChatGPT Work, AlphaEvolve, humanoid robots at BMW. This week was about finishing, not assisting.
- Read
What Liberty Mutual knew that most companies didn't
When Fable 5 went dark, Liberty Mutual kept running. Their infrastructure approach is the playbook every operator should know.
- Read
The $2B bet that AI wins through acquisition
SoftBank and Thrive are buying accounting and IT firms to transform them with AI. Plus four other reads for today.
- Read
A 35B model just matched a 1T model on agent tasks
The model commodity crash arrived quietly last Monday. Here is what it means for operators trying to build something durable.
- Read
The government just became an AI counterparty
The government pulled a model, got an equity offer, and Anthropic started building its own chips. All this week.
- Read
Fable 5 is back. Coding costs just went up.
Fable 5 returns globally today with a forced fallback to Opus 4.8 for coding tasks. Plus four other moves that matter for operators this week.
- Read
The AI success rate your team is probably ignoring
Spotify's AI workflow isn't about more Claude sessions; it's about the verification layer they built around them.
- Read
The AI Consultants Are Being Disrupted
Accenture just lost 20% in a day. Here's what that says about whether you still need a consultant
- Read
A $30M startup wants to watch every AI agent in your company
Claude moved into Slack. Gong shipped a governed-agent layer for revenue teams. Runlayer raised $30M to watch them.
Get every issue, as it lands.
What changed in AI this week, and what it means for the work your company does.