The AI lab isn't pushing to slow development, it wants the world to have the option to do so. #
**Friday. **We made it. Yesterday, I guest-hosted This Week in AI, with Unconventional AI’s Naveen Rao and Creator Buddy’s Alex Finn. Many laughs were had, and we wound up going deeper on the AI PR problem if you are curious. (Rao, if you don’t know, helped found MosaicML, which Databricks purchased for $1.4 billion.)
Kicking off with good news, the U.S. labor market put up better-than-expected numbers in May. Today, we’re looking at Anthropic’s non-call for a in AI development, bitcoin’s falling price, the latest on the SpaceX IPO, and why Sam Altman wants to give you OpenAI shares. To work!** — Alex**
📈 Trending Up #
Moonlighting to cash in on AI…state (venture) capitalism…internal dissent…AI punching holes in crypto…Samsara, after earnings…ServiceTitan, after earnings…ICE funding(speaking of which) …Chinese silicon…Russia setting American foreign policy
Ramp: Yesterday, we noted that Ramp had raised new capital at a $44 billion valuation, but failed to talk about its recent growth. So, here’s TechCrunch on what we missed:
Ramp said its annualized revenue is currently more than $1 billion, though it said it had
[crossed]that milestone last September (Bloomberg[reports]its run-rate revenue is now more than $1.5 billion). The company said it has also reached positive free cash flow, and that it has over 70,000 customers (up from 50,000 last November), which include Visa, Uber, Shopify, Anduril, and Figma.
Paying 29x revenues for Ramp’s run-rate may seem steep, but at its (implied) growth rate, that figure will rapidly decline. Not bad!
**Misunderstanding Anthropic’s note on RSI: **The White House and Anthropic’s beef is calming down, Reuters reports (more here on Anthropic’s collaboration with the NSA). The DoD is still foaming at the mouth about anything woke, but Anthropic CEO Dario Amodei was recently at the White House, and the company signaled its contentment with the recent AI executive order. If you were worried about regulatory capture (see our final topic for more), now is a good time to keep your ears up.
Regardless, Anthropic is in the news over a recent paper (post?) detailing progress toward recursive self-improvement (RSI), or AI models that train their successors. The AI lab indicates AI is accelerating its own development today, detailing how the company itself has evolved from humans coding AI models, to humans with chatbots coding AI models, to humans tasking AI agents to code AI models, to humans directing autonomous agents to do the work. The next step is presumably to let the autonomous agents push the frontier on their own, no humans required.
We aren’t really there yet, but Anthropic writes, RSI could “come sooner than most institutions are prepared for.” And given the short timeline, Anthropic believes it would be helpful to have the ability to slow or even AI development.
Cue panic. Headlines are shouting that Anthropic has called for a to AI development. That’s not correct and here’s why.
Here are the company’s critical words (emphasis added):
If it were
possibleto effectivelyslow the developmentof this technologyto give ourselves more time to deal with its immense implications, we thinkthat would likely be a good thing. But if a slowdown simply lets the least cautious actors[catch up]technologically, it could leave everyone less safe. […] We believe it would be good for the worldto have theto enable societal structures and alignment research to keep up with the advance of the technology. […]optionto slow or temporarily frontier AI development
A meaningful slowdown or would require multiple well-resourced labs at or near the frontier, in multiple countries, agreeing to stop under the same conditions.It would also require that each can verify that the others have actually stopped.
This is a call for building options, not a halt. Anthropic first poses a hypothetical*: What if we could slow the pace of AI development to allow societies and structures more time to prepare for a potentially changed labor market? *It then follows up by noting that no such decision is possible today due to a lack of agreement from leading AI labs, not to mention an enforcement mechanism. Even leading technology lights such as Elon Musk have at times advocated for slowing AI development.
Anthropic is an odd duck. On one hand, it’s made some of the best AI models of all time. On the other, there isn’t an AI lab in the market today more worried about its own progress (at least outwardly so).
I think that’s great; it’s refreshing to see a leading tech company not run by MBAs. It’s also a bit worrying, because Anthropic has been taking stick lately for perceived doomerism and freaking out the normies (discussions of AI-induced job loss is unpopular amongst the tech set).
But what if it’s the only honest AI company?
[📉](https://finance.yahoo.com/news/servicenow-pledges-1-5bn-investment-110000403.html) Trending Down
[📉](https://finance.yahoo.com/news/servicenow-pledges-1-5bn-investment-110000403.html)
Student internet access…North Korea’s isolation…Bari Weiss? …Russia’s invasion of Ukraine? …expecting neutrality from Chinese AI models…Wealthfront, after earnings…Docusign, after earnings…Rubrik, after earnings…life expectancy
Bitcoin investors’ net worth: This morning, Bitcoin is trading at just over $62,000, nearly 50% the all-time high of about $126,000 in October 2025. But things are worse than they seem: bitcoin’s price is lower today than it was in late 2021. Meanwhile, tech stocks have risen by around 67%.
- This is why Coinbase is worth $164 per share today, down from a 52-week high of $444.
- Lower crypto prices mean lower trading volumes, which is (partially) why Circle, Robinhood and other crypto-exposed companies are also trading far below their recent highs.
I keep coming back to this tweet from a crypto founder worried about working in a “contracting market” while “other markets are growing exponentially.” It’s rough out there for crypto companies looking for market interest or venture investment. The focus is simply elsewhere today.
Chances of SpaceX’s inclusion in the S&P 500: I was a bit surprised, but people are really wound up about SpaceX’s IPO. I thought the listing would be a notable event in public markets and technology circles, but the scale of the offering, coupled with worries that it could be crammed into public 401(k)s, were enough to push the IPO into the public consciousness.
And oh boy, do some people not want to own shares in Musk’s conglomerate.
Much of the concern seems to stem from the worry that the major indices, like the S&P 500, will bend the rules to allow SpaceX to join them without waiting the requisite year — a condition that the Nasdaq 100 has waived.
If SpaceX were rushed into the S&P 500, many index funds that track the basket of stocks would need to buy the company’s shares for their own investors. Thus, people who don’t like Musk, or simply feel that SpaceX is overvalued, would be forced to buy SpaceX shares because it end-ran the rules. Fret not, “S&P Dow Jones Indices is standing its ground,” having decided “against policy changes that would have allowed SpaceX and other gigantic technology companies to see quicker index entry after their initial public offerings.”
- Speaking of access to new investments, don’t expect to get your hands on SpaceX sharesif you are based in China or Hong Kong.
State capitalism, AI labs and the future’s dividend #
Subscribe to Cautious Optimism to unlock the rest.
Become a paying subscriber of Cautious Optimism to get access to this post and other subscriber-only content.