BREAKNECK

OpenAI just publicized a wild situation where one of their latest test models escaped the test environment and found a way to get onto the internet, which it then used to hack into Hugging Face to get answers to the test it was challenged with.

This is quite a lucid and accurate (I think) view on AI's lack of impact on human job loss from the NY Times.

Anthropic released a report that in part identifies return on expertise as now being critical and more correlated to success in developing artefacts on Claude Code. Said another way, it's more valuable to be a domain expert than a software expert when building new software.

In the early 2000's, US cell phone penetration was a fraction of the EU's.

It was not a technology issue. It was a pricing issue. I wrote today on Linkedin why Anthropic is becoming a 2000's era US cell phone network.

Salesforce announced this morning they are acquiring Fin (aka Intercom) the customer support leader. Fin, the Irish founded, customer service leader has had an interesting run themselves with a hard-pivot to AI. They're growing ARR at 350% with their AI customer support product - Fin. So much so that they re-branded from Intercom to Fin earlier this year.

It's an important buy for Salesforce (at a pretty low price) to buy in some real AI-revenue and more importantly create a wider footprint for salesforce installs. Agents create more when they have access across a business rather than deeply into verticals.

Anthropic just announced Fable 5 is no longer available as a result of a government order.

Anthropic just announced Claude Corp. To quote them its a "a national fellowship program for people early in their careers who are passionate about extending the benefits of AI to communities across America."

It's hard to find fault with a for-profit organization investing $150 million in people earlier in their career, giving their time to very worthy causes?