Damn, that looks awful.

Where can we read the code?

> if it’s a solo greenfield project

That's a big if. I don't have numbers but most professional engineers are not working on such projects

I've been thinking of doing the exact same thing. Preserve context as images and die. Expose a single tool called "eval". You could have a extremely tight editor integration using something like SLIME.

Still nothing to compete with GPT-OSS-20B for local image with 16 VRAM.

> When something breaks, I SSH in, ask the agent what is wrong, and fix it.

> I am spending time using software, learning

What are you actually learning?

PSA: OP is a CEO of an AI company

you are learning what it takes to keep a machine up and running. You still witness the breakage. You can still watch the fix. You can review what happened. What you are implying from your question is that compared to doing things without AI, you are learning less (or perhaps you believe nothing). You definitely are learning less about mucking around in linux. But, if the alternative was not ever running a linux machine at all because you didn't want to deal with running it, you are learning infinitely more.

How can you review if you don‘t know in the first place?

You can watch your doctor, your plumber, your car mechanic and still wouldn’t know if they di something wrong if you don’t know the subject as such.

You can learn a lot from watching your doctor, plumber or mechanic work, and you could learn even more if you could ask them questions for hours without making them mad.

You learn less from watching a faux-doctor, faux-plumber, faux-mechanic and learn even less by engaging in their hallucinations without a level horizon for reference.

Bob the Builder doesn't convey much about drainage needs for foundations and few children think to ask. Who knows how AI-Bob might respond.

> You can learn a lot from watching your doctor [...] work

Very true but I'll still opt for that general anesthesia...

The primary way humans learn anything at all is by watching and mimicking. Sure, there will be mistakes, but that doesn't preclude learning.

Your hypothetical situation would cause all progress to halt. Nobody would be able to fix genuine problems.

who cares if I'm the CEO of an AI company? I didn't mention anything related to my company once in the post.

Wrote about learning and fun here: https://fulghum.io/fun2

It's (at least) common courtesy to declare potential conflicts of interest.

I would have if there were potential conflicts.

I personally find this information quite relevant to get the full picture and I'm glad someone else provided it.

The full picture of what exactly? How that fact is even relevant to this post? Do you expect anyone affiliated with AI to mention that every time they talk about AI? That's just ridiculous.

I expect someone writing a blog about AI agents help you run your home server to disclose that they are "helping companies automate operations with AI" as their job, which they get money for.

Why wouldn't you bring it up, or even lead with it?

Doesn't it make sense to want to know this? It's not far fetched at all that there is a conflict of interest. How can they be unbiased in the validity of the approach if this is exactly the same stuff they sell for money?

If international law had any effect people would believe in it. You're mixing cause/effect. This situation has been going on for years and the lack of response by international organizations makes people lose all confidence in them.

It's the same with money: if people believe in it it works, if they stop believing in it it stops working.

So there is no cause and no effect, it is something mutually reinforcing.

I'm unfortunately on the same situation. We made a consultation with people from Baptist Health Miami and it seems like there are several non trivial requirements for such treatment (histotripsy), like the number and location of mets. Hope that this improves in the mear future.

One of the first things I tried in Rust a couple of years ago coming from Haskell. Unfortunately it's still not stabilized :(

> Memory safety is not that sacred. In fact, for many applications malfunctioning is better than crashing

Yeah, just another data leak, no biggie.

What's a bit of arbitrary code execution between friends?