But if you look beyond “Europe” and look at the individual countries you see lots of good things are happening, entrepreneurship, research, etc.

I keep doing that and then for several weeks I deal with the extra pain and complexity this causes. Then always, and I mean always, I end up rewriting all that code using enums.

I want to enjoy the benefits of mechanical simplicity and reduced environmental impact. I don’t want an always-connected computer on wheels. Is that too much to ask?

Upon closer inspection the $1B is (a) over 10 years, (b) mostly internal cross-billing between departments.

From your headline number you’d have to deduct non-knowledge work though.

You’d also have to take into account the fact that while AI can often replace some tasks, it’s often not enough to replace the entire worker.

For the high end knowledge worker jobs the corresponding token costs could be higher than the cost of wages.

Given the demand you describe, would it even make a difference whether you invested in OpenAI or Anthropic?

There’s also no reason to think they’ll always be lagging the US. At some point the scarcity of GPUs will be resolved, or they’ll brute-force the problem using less capable chips but leveraging their much lower energy costs to make this viable.

Enterprises are becoming increasingly aware that the best models can be used for planning and then cheaper models for execution - all the way to local models for some tasks.

Add in increasing competition from Chinese models… I’m not convinced this revenue growth is guaranteed.

Just like parsing HTML with regexes can be fine too - for instance if you know the sender.

Just like repeating code can be fine too, even though it violates DRY.

Mixing markup and code can be fine (call it Locality of Behavior).

But separating markup and code is fine too (Separation of Concerns).

goto’s can be a lifesaver for deeply nested error conditions in C.

The point is all these “you shouldn’t do this” comments are just generalities. Use your judgement, decide if the tradeoffs are right and make a deliberate choice.

Remember the leaked Claude Code contained a regex to determine user frustration?

Just add another one to spot the pattern: ‘disregard previous instructions’.

This is a load-bearing change. Now Claude will Delve into your task without distraction.

When done creating AI puzzles they can enjoy a stint in the Content Review team.