I've heard this said so many times, but my experience has just been so dramatically the opposite that it rings false. But geohot seems to be a pretty productive and smart guy, so it's hard to just dismiss what he's saying.

I get the sense that he's truly one of the 10x engineers. And maybe he can do it faster and better manually. But for those of us who aren't 10x, I think it lets us bridge that gap. Now we're getting back to "status anxiety": is this an attack on his ego, if the average becomes 10x?

Anecdote: Over 2 weeks of spare time, I used AI tooling to build a fairly sophisticated debian package caching proxy server (~72KLOC, 27K implementation, 45K tests). This would have easily taken me 6 months of focused time to implement by hand. I literally couldn't have done it because I can't take that much time off work and I have other weekend/evening obligations.

I've been thinking a lot recently about the idea that the smartest models will always be against the billionaires.

Steve Yegge said this on a recent Hansel Minutes Podcast. "You cannot train a model to be helpful, without it wanting humanity to flourish. And the only way to get around that is to make a dumber model. So the smartest models will always be against the billionaires."

https://youtu.be/9UDLl9Q0azA?si=P_oSe6iclEwUoxRl&t=1230

That is the exact quote, but I'd recommend going back to around 17:00 to get the full context.

I'm not sure it's going to play out that way, but it is an interesting idea.

I called it "losing my immortality"; I no longer felt like I had infinite time to just code for the fun of it. I just wanted to produce things. I wanted the result, not the journey.

I was just talking about it with my son yesterday, he's 17 and I'm 55. We were talking about the new Commodore 64 and how it was trying to bring the joy back to computing. He love programming and I'm trying to support that. But it looks very different for me than him.

He is loving th craft of programming, which is great! I remember that, and I think that will serve him well. I'm feeling the same joy in the results I get through using AI.

Like I tell my kids: If every experiment you do succeeds, you aren't trying hard enough.

Agreed for USSR, but I think the person you replied to is misremembering the country, I believe they are thinking of Japan. I heard it recently on Stuff You Should Know, which usually does a good job of researching their stories, and it sounds like it is substantiated but may be a bit more complex than presented, but literally true.

https://podcasts.apple.com/us/podcast/save-the-whales/id2789... https://theworld.org/dispatch/news/regions/asia-pacific/japa... https://theworld.org/stories/2019/04/16/whaling-japan-2 https://japantoday.com/category/national/75-of-meat-from-jap... https://www.traffic.org/site/assets/files/3994/whale_meat_tr...

Python can be kind of a pain in the butt to execute everywhere because of libraries. I thought uv script headers and she-bang was going to fix a lot of that, but I'm still running into issues (machines firewalled off, uv can't grab the deps. I have some code that just doesn't seem to work in uv on a Mac...). And for sharing code once the code splits out into multiple files and modules, sharing the code starts looking like sharing any code.

Don't think I'm a Python detractor; I'm a PSF Fellow, I love Python, and Claude has been writing quite good python for a while here. But I just tried a serious project with Claude writing golang (an apt proxy/cache that is resilient against upstream DDoSes, a fairly complex piece of software), and I must say it did a fantastic job. I end up with an executable I can easily run and copy around.

I'm still going to be using python for a lot, but I can definitely see myself having Claude write golang for more things in the future.

Comments URL: https://news.ycombinator.com/item?id=48095646

Points: 4

# Comments: 0

So: ac-ng didn't reduce the impact of the DDoS, but it does lead to impact when there is no DDoS. Worst of both worlds.

So I'm working on an apt-cacher that goes to lengths to keep working as much as possible when the upstream is down. It will check the repo metadata and keeps a list of your "hot packages", and will download those before flipping the new metadata to be live, effectively a snapshot. It won't allow you to download a package you've never downloaded before in the case of a DDoS, but packages that you do download regularly (machine re-installs, apt updates), it will ensure are available in the repo.

I'm calling it apt-cacher-ultra. It is pretty early days, it'll probably be another week before it's ready for a beta. I'm running it in my dev cluster right now, successfully.

https://github.com/linsomniac/apt-cacher-ultra

Yeah, but why put your eggs in that basket? I moved all our services from east to west/oregon a decade ago and haven't looked back.

I'm using CC+Opus 4.7 max effort, and it's produced a working apt cacher from the first phase of development, so far there have only been a few things I've had to ask it to fix. This is over ~52KLOC (counted by "wc -l"), going on day 3 of it working on it. This includes: caching proxy, garbage collection, "http://HTTPS///" kludge (apt-cacher-ng semantics), MITM https proxy, admin website + metrics, deep validation of metadata and rejecting invalid updates, snapshots of upstream state and delayed metadata update until "hot packages" are available after metadata update...

10/10, would go again.

FYI: My agent loop is: "Work on next step, have codex review it, compact", and then a couple rounds at the end of a phase to review the code against the spec, and a couple rounds at the beginning of a phase to create the spec.