This Kills OpenClaw
This Week in AI: Claude's ship-a-thon. Sora's funeral.
Hello Futurists,
This week Anthropic pushed Claude into full computer control, OpenAI killed off Sora, and Google dropped a breakthrough that could slash AI costs and speed everything up…
Let’s get into the news!
- Ejaaz
Claude Gets Computer Control (This Kills OpenClaw)
Anthropic has been on an absolute tear this past couple of weeks — launching 9 new features in 10 days.
The most significant was Computer Use — Claude can now access any app, tool, desktop or browser and intelligently operate it… like a human would. Any computer-based task, Claude can now do.
This is huge for a few reasons:
OpenClaw, the open-source personal agent, went viral for being able to do this, however the product itself was clunky and exposed you to major risk. This new Anthropic product is much safer and arguably better.
Claude has the ability to automate white-collar work. It’s not quite there in V1, but this is where it all gets started.
Insane execution speed from the Anthropic team.
OpenAI Kills Sora
OpenAI has shut down their AI video generation app Sora, the app made famous for its TikTok-style video reels that reached #1 on the App Store just days after launch.
OpenAI cites the reason for shutting it down as consolidating work resources, compute and money to focus on building better AI models and coding agents.
Sora had a pretty short life! It only launched six months ago, but achieved some impressive feats such as 1 million downloads in the first 24 hours of launch — and consuming 30% of OpenAI’s entire compute capacity on a given day.
Hollywood seems to be rejoicing at the news with plenty having called the app an “AI slop” generator.
Google Popped the AI Bubble
Okay, not exactly, but we’re close… they invented a new AI algorithm that shrinks the memory requirement of an AI model by 6X… without killing its intelligence.
That means frontier AI models that cost $100Ms to train might actually be a lot cheaper if this algorithm can scale. The research shows models also become 8X faster. It gets better: this uses the same number of GPUs which means all the expensive GPUs companies are spending billions on might get pushed into turbo-mode.
OpenAI Is ‘Guaranteeing’ Returns
OpenAI is using a bold strategy to win back the enterprise market from Anthropic.
They’re offering private equity firms a guaranteed 17.5% return if they decide to roll out ChatGPT into their portfolio companies (which they own). The idea would be ChatGPT gets installed into these enterprise companies and OpenAI’s product helps them make better profits and revenue using AI.
Many are viewing this move negatively, citing that OpenAI might be getting desperate — Anthropic currently holds a massive advantage among first-time enterprise users, with 73% opting for Claude.
SpaceX Filing for IPO Next Week?
The rumors might be true… SpaceX is reportedly filing to go public as soon as next week. This would prime them for a June IPO.
The rumors also state they’re seeking to raise a staggering $75 billion and launch at a $1.75 trillion valuation. If true, this will easily be the biggest IPO of 2026, even ahead of OpenAI and Anthropic’s blockbuster launches.
Thanks for joining us for another issue. Now go listen to our podcast :)
But your wrong by saying anything... what it cant do is anything financial. So it cant trade for you. Just saying.
Sharp framing. The part teams underestimate is not model quality, it’s operational drift after week 2. One thing that helped us: after every automation run, log 5 lines (goal, tool calls, failure, recovery, next test). It makes “this kills X” claims falsifiable fast. If useful, Giving Lab shares practical examples of that loop from real OpenClaw workflows here: https://substack.com/@givinglab