What’s happening in the world of AI:

Highlights

Anthropic Takes the Pentagon to Court

Discussed at [01:47]

Anthropic is are preparing to legally challenging the supply chain risk designation. They're calling it unprecedented, legally unsound, and a dangerous precedent for any American company that negotiates with the government. Legal experts reckon the Pentagon’s decision will fall apart in court. But…that could take months, and the commercial damage gets done in the meantime.

The tech industry is rattled. An open letter with 56 CEOs and 47 OpenAI employees is demanding the government backs down. Even Anthropic's biggest rival thinks this has gone too far. Because if the government can blacklist any domestic company that says no, every founder in America just became expendable.

Meanwhile, Amazon is hedging. No-one from Amazon has signed the open letter tellingly. And there hasn’t been a statement from Amazon about the whole situation.

They've been Anthropic's biggest backer but just put money into OpenAI's $110 billion round. The largest private funding round in history, by the way. Previous record? $40 billion. Also OpenAI. Third largest? $30 billion. Anthropic, last month. The top three VC deals ever have all happened in the last twelve months…

Kyle's take: I feel that the six months Trump has given Anthropic is a TikTok-style get-out. The deadline will keep sliding. A deal gets done behind doors. Maybe the investors push for Dario to be replaced with someone more compliant. Though Anthropic without Dario? Can't even imagine it.

Honestly, I think a lot of this is more basic than the ethics debates suggest. Dario is high-minded and a bit heavy-handed. I reckon he just rubbed the Pentagon the wrong way. OpenAI navigated the same waters with basically the same red lines because Sam Altman is smoother. Pure speculation of course!

Source: The Hill | Tech workers open letter | Open letter signatures

ChatGPT Uninstalls Surge 295%. Sam Backpedals.

Discussed at [14:57]

The numbers are brutal. ChatGPT uninstalls up 295%. One-star reviews up 775% in a single day. Downloads down 13%. All because OpenAI swooped in on Anthropic's Pentagon contract hours after Trump's ban. The body wasn't even cold guys!!

Claude hit number one on the US App Store. Katy Perry signed up(?). The #QuitChatGPT movement is going viral. For a company that nobody outside tech had heard of two weeks ago suddenly Anthropic are a big deal and OpenAI look like the “baddies”.

Sam's backpedaling now. Defending the decision while admitting the optics weren't great. He's in an impossible spot: defender and vulture at the same time.

But let's keep the Anthropic hero narrative in check, eh? This isn’t a simple good guys vs. bad guys. Anthropic still took a $200 million Pentagon contract. They knew what they were getting into.

And read Dario’s statement carefully. They are fine with kill bots, just not ones that might harm Americans - the tech isn’t ready yet and may harm US warfighters. Still fine with surveillance, just not domestic surveillance. Turn those autonomous weapons on the Middle East and surveillance on China? Cool, no problem.

They're the good guys in this specific fight. They're not pacifists. So let’s not go too wild here.

Kyle's take: I've been beating the Claude drum for two years and nobody listened. Now everyone's angry at ChatGPT and suddenly interested. Welcome, I guess! But don't cancel ChatGPT out of pure rage and have nothing. Try Claude first. Use both for a week. Decide with your head.

Source: ChatGPT uninstalls 295% | Sam Altman defends deal | CNBC amended deal | Uninstall data

 Claude Is Down. And It's Not Vibe Coding's Fault.

Discussed at [19:01]

Claude's been struggling for days under its own popularity. Elevated errors on Opus 4.6, login issues, the lot. The conspiracy theories are rife: Iranian missiles hitting a data centre, Pentagon DDoS attacks, vibe coding breaking everything. The actual reason? Too many people downloaded it at once. They just weren't ready.

The smart move: they’ve now made that the ChatGPT memory import tool free. It was paid-only initially, which was stupid…make migration from ChatGPT as easy as possible Now anyone fleeing ChatGPT can export memories and import them into Claude without spending a penny.

Kyle's take: Anthropic need to fix this yesterday. Every hour of downtime is thousands of converts who'll never come back. But there's also the massive question of whether AWS will keep hosting a blacklisted company when Amazon has its own government contracts to protect… and Amazon just invested into OpenAI. It’s getting messy!

Source: Claude outage | Claude status | Memory update

iMember Question from audience: "Should I cancel ChatGPT and switch to Claude?"

Kyle's response: I've been saying use Claude for two plus years!! Go and try it. Give it a task ChatGPT struggled with. Claude will probably blitz through it. But be realistic: servers are hammered, your first experience might be a loading screen. Be patient. And don't cancel out of rage and have nothing. Use both for a week. Then decide.

Also be aware that Claude limits are much stricter than ChatGPT. Very annoying!

This question was discussed at [22:52] during the live session.

Member Question from audience: "What niche should I teach AI in?"

Kyle's response: Whatever you already know. The job isn't technical, it's translation. You're bridging AI and the people in your industry. That requires stories, metaphors, language that makes sense to them. I charge about $2,000 an hour for workshops and I don't have a computer science degree. I have a history degree. It's about communication, not credentials. The best niche is the one you already live in.

(Related: I’m giving a webinar tomorrow on teaching AI to businesses - free sign up )

This question was discussed at [53:50] during the live session.

Streaming on YouTube (with full 4k screen share) and TikTok (follow and turn on notifications for Live Notification).

Audio Podcast on iTunes and Spotify.