Hello AI Enthusiast,
This week, the US government forced Anthropic to pull its most powerful models off the market, SpaceX spent its fresh IPO capital on an AI coding tool, and the EU gave companies a roadmap for labelling AI-generated content. From government intervention to billion-dollar bets, let's dive in.
The Big Picture 🔊
US Government Orders Anthropic to Pull Its Most Powerful Models
The US government issued an export control directive requiring Anthropic to suspend access to Fable 5 and Mythos 5 for all foreign nationals, inside or outside the US. To comply, Anthropic had to disable both models for all customers. The government cited a jailbreak method, but Anthropic reviewed it and found the same capability is already available in other publicly deployed models. All other Claude models remain available.
|
||
|
This was predictable in a way. Anthropic spent months warning that Mythos was powerful enough to raise serious security concerns, and the US government took that seriously, just not in the direction Anthropic hoped. When you build your own case for why a technology needs special controls, you hand regulators the argument to use against you. What makes this interesting is the broader dynamic at play. AI is the first major disruptive technology to come from private companies rather than government-funded research. Nuclear, aerospace, the internet: all had government at the center from the start. With AI, governments are playing catch-up, and this directive feels like one of those catch-up moments. Whether it holds up to technical scrutiny is a separate question. |
SpaceX Buys AI Coding Tool Cursor for $60 Billion
Days after its record IPO, SpaceX agreed to acquire Cursor, a popular AI coding tool with over $1 billion in annualized revenue, in an all-stock deal worth $60 billion. The move follows SpaceX's merger with xAI earlier this year, positioning it to compete with Anthropic and OpenAI in the AI coding space. Notably, Cursor's market share had dropped from 41% to 26% in under a year before the deal.
|
||
|
The timing here says a lot. SpaceX just had the biggest IPO in history, and within days it is spending that capital on an AI coding tool. Grok never really took off as a standalone AI product, and this looks like a second attempt to compete with OpenAI and Anthropic, this time with a tool that developers actually use. |
EU Publishes Code of Practice for Labelling AI-Generated Content
The European Commission published a voluntary Code of Practice ahead of AI Act transparency rules taking effect on August 2nd. It requires providers to mark AI-generated content in machine-readable formats, and deployers to label deepfakes and AI-generated public interest text that has not undergone human review. Signing the Code simplifies compliance across all EU member states. Companies that don't sign still have to comply with the Act, just through other means.
|
||
|
This is voluntary in name only. The AI Act transparency obligations kick in on August 2nd regardless. Signing the Code just makes compliance easier to demonstrate. Companies that don't sign still have to comply, just with more paperwork. The labelling requirement itself makes sense for deepfakes and manipulated media, where the stakes are clear. Where it gets complicated is with AI-generated text. If a journalist uses AI to draft an article and then edits and reviews it, no label is needed. If a news site automates publishing without human review, it must be labelled. |
Bits and Bobs 🗞️
Microsoft's Copilot Cowork, now generally available, runs complex multi-tool AI tasks with improved cost efficiency, user controls, and enterprise-grade security using Anthropic and upcoming Cowork 1 models.
OpenAI now lets users save Codex rate limit resets to use later, starting with one free reset for Go, Plus, Pro, and Business users.
Anthropic Public Record survey found broad agreement that Americans want AI benefits like curing disease but fear job loss and cognitive dependency, with strong bipartisan support for government regulation and company accountability.
OpenAI has launched the OpenAI Partner Network, investing $150 million to help global partners build and deploy AI solutions that integrate OpenAI models into enterprise workflows at scale.
Google's Gemini AI will soon integrate with Google Business Profile to provide small businesses with personalized, context-aware assistance for managing reviews, data, and workflows.
Meta's new AI Mode on Facebook allows users to ask questions in plain language and get synthesized answers from public posts across the platform.
Google's Pixel Drop update introduces AI-powered Screen Reactions for recording selfie videos with screen content, voice translation, and AI-driven Magic Cue suggestions in chats.
We're hiring!🚀
We're looking for an AI Solutions Consultant / Automation Engineer to join the team. If you build real automations and know how to run B2B projects end to end, we want to hear from you.
From Our Founder’s Channels 🤳
In his recent Linedin post, Gianluca Mauro talks about what CEO’s worry about when it comes to AI: how to create an operating model to embed AI in the organization. Read it here.
Most teams already have AI. Far fewer get real value from it. Our Corporate AI Training gives your people the skills and habits to turn AI from a license you pay for into results you can see.
LOLgorithms 😂
Manifesting gone wrong.
That's a wrap on our newsletter! Before you go, here’s a quick recap of our offerings:
AI Academy Membership: Get 12 months of access to all our cohort-based programs and on-demand courses.
AI Agent Bootcamp: Accelerate processes and solve business problems by building AI Agents, without coding.
Practical Introduction to ChatGPT: A free course on using ChatGPT confidently, understanding its workings, and exploring its potential.
Customized Corporate Training: Equip your team with the skills they need to unlock the potential of AI in your business.
Catch you next week! 👋
