I suppose it’s one thing to rewrite code and make tests pass. But a lot of my own CPU cycles when coding go to making the code understandable, modular, editable etc. IME Claude is not great on these.

Maybe it doesn’t matter? Maybe the spaghetti makes sense to Claude, and when you say “hey add this feature”, no problem?

But maybe it will turn into a worse ball of spaghetti, with no nicely curated tests, hacks on hacks on hacks and never ending loops of whackamole of “just change this one line and rerun tests”.

I can imagine both - let’s see.

But the real sales pitch is that AI overtakes everything. That the YC cohort of 2032 will be just CEO, sales guy and a massive AI bill. It’s not entirely impossible IMO, either.

Then you totally get network effects. All your company documents, discussions, context etc are in there, your agents/employees whom you finally taught to do the job right. And I’m guessing the REST API for extracting your data is absent.

I can’t find any handy data, but my understanding is that US and China certainly have the most fossil fuel in their new power plants. As well as, notably, China putting online a lot of renewables and nuclear in absolute terms.

Not much of a consolation but I want to give credit where it’s due.

Europe is also potentially more affected by climate change than US and China, as I understand. India of course would be toast in a hotter world, but perhaps it too poor right now to do much about it easily.

But discoordinated action won’t move the dial. So should Europe continue on the green path while losing GDP share to US, China and India? Should it jump on the coal bandwagon and screw up the world since everyone else is? It’s not like green contrarian action will do much good on its own.

It’s altogether a not pretty picture…

The GFC “proper” was the dramatic crash in the liquidity of credit markets, not strictly a corollary of the losses on property and mortgage-backed securities.

What does make me upset is that Europe has the technology and seems unwilling or unable to get it's act together and make some chips. China seems to have managed this part too, only without access to the latest chip making technology!

It is also true that many times the Polish state has behaved inhumanely towards these people. But for sure their presence on the forester Polish-Belarussian border was an engineered hybrid warfare act, not some act of collective desperation, and it continues to this day.

The poor people taking part are, I understand, often recruited in poor places and told they will be smuggled into Europe. Then they arrive in Belarus and get sandwiched between Belarusian police (KGB?) and the border, told the only way out of, effectively, Belarusian captivity, is into Poland, through the barbed wire and border patrol/army.

The issue with emulating Emacs is that the editor key bindings are not the feature. The feature of Emacs is basically an IDE totally tuned to how you want it, plus the fact that then you apply the same ecosystem to a great number of text-adjacent applications - Org, E-mail, code, markdown, etc.

Someone saying to me "hey, you can copy with M-w!" just really doesn't appeal.

Vim seems more about an out-of-the-box "language" for typing, so I can see how you can emulate this more sanely.

LLM code past a few pages is really unreadable. A workflow of LLM writes code, human reads/checks/corrects is IME very unproductive and frustrating, and you're usually better off writing by hand (YMMV).

If you want to reap some productivity improvements, it can't rely on you reading, grokkong and accepting the LLM code line by line. You just kind of need to suck it in.

So the best you can do is to indeed, constrain it heavily, maybe write the interface, then have the LLM write tests, then have it write the code.

I'm not saying I like the approach, but I agree manually verifying LLM code is very frustrating and unproductive.

As a plus, it would have likely left the US in a better place than having fought these wars!