My AI skeptic friends are all nuts

“I work mostly in Go.”

That should have been the start of the essay.

New title: LLMs make Go programming feel good

How are you writing your prompts? I usually break a feature down to smaller task level before I prompt an agent (claude code in my case) to do anything. Feature level is often too hard to prompt and specify in enough detail for it to get right.

So I'd say claude 4 agents today are at smart but fresh intern level of autonomy. You still have to do the high level planning and task break down, but it can execute on tasks (say requiring 10 - 200 lines of code excluding tests). Any asking it to write much more code (200+ lines) often require a lot of follow ups and disappointment.

I hate how the discourse around LLM-assisted programing is so polarized. It's either detractors saying it's "a fad that's useless and going nowhere, wasting billions of megawatts every year" or it's true believers calling it "the most miraculous sea change technological advancement in my lifetime" or "more important than fire and electricity[1]." There just doesn't seem to be any room in the middle.

I tried out Copilot a few months back to see what all the fuss was about and so that I could credibly engage with discussions having actually used the technology. I'd rate it as "kind of neat-o" but not earth shattering. It was like the first time I used an IDE with auto-complete. Oh, cool, nice feature. Would I pay monthly for it? No way. Would I integrate it into my development workflow if it were free? Maybe, I guess? Probably wouldn't bother unless it came literally set up for me out of the box like autocomplete does nowadays.

Don't get me wrong--it's cool technology. Well done, AI people. Is it "the 2nd most important thing to happen over the course of my career" as OP wrote? Come on, let's come down to earth a little.

1: https://www.cnbc.com/2018/02/01/google-ceo-sundar-pichai-ai-...

> The author has some valid points, but dismissing this entire class of arguments so flippantly is intellectually lazy.

Agree 100%. And generally programmers have a poor understanding of the law, especially common law as it applies in America (the country whose legal system most software licenses have been written to integrate with, especially copyleft principles).

American Common Law is an institution and continuity of practice dating back centuries. Everything written by jurists within that tradition, while highly technical, is nonetheless targeted at human readers who are expected to apply common sense and good faith in reading. Where programmers declare something in law insufficiently specified or technically a loophole, the answer is largely: this was written for humans to interpret using human reason, not for computers to compile using limited, literal algorithms.

Codes of law are not computer code and do not behave like computer code.

And following the latest AI boom, here is what the bust will look like:

1. Corporations and the state use AI models and tools in a collective attempt to obfuscate, diffuse, and avoid accountability. This responsibility two-step is happening now.

2. When bad things happen (e.g. a self-driving car kills someone, predictive algorithms result in discriminatory policy, vibe coding results in data leaks and/or cyberattacks), there will be litigation that follows the bad things.

3. The judges overseeing the litigation will not accept that AI has somehow magically diffused and obfuscated all liability out of existence. They will look at the parties at hand, look at relevant precedents, pick out accountable humans, and fine them or---if the bad is bad enough---throw them in cages.

4. Other companies will then look at the fines and the caged humans, and will roll back their AI tools in a panic while they re-discover the humans they need to make accountable, and in so doing fill those humans back in on all the details they pawned off on AI tools.

The AI tools will survive, but in a role that is circumscribed by human accountability. This is how common law has worked for centuries. Most of the strange technicalities of our legal system are in fact immune reactions to attempts made by humans across the centuries to avoid accountability or exploit the system. The law may not be fast, but it will grow an immune response to AI tools and life will go on.

Have it make small changes. Restrict it to a single file and scope it to <50 lines or so. Enough that you can easily digest without making it a chore.