I freaking fell for that, like an idiot. I distinctly remember a conversation with my brother in law who is a truck driver who literally guffawed when I told him about all the AI driving research and the prospect of self driving trucks.

I learned my lesson. Fool me once...

Name names you stupid fucker if you're going to say shit like that. Who is making $10 million fucking around with LLMs to make products for customers? Tell me so I can avoid them.

I use LLMs a *lot* in my work. I've been programming since 1989, and the LLMs are certainly useful, but saying you can build software with "vibes" is an incompetent level of understanding of what makes software that is morally acceptable to sell to customers.

Get an LLM to replace Garry Tan, because for sure that wouldn't be worse.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fdaringfireball.net%2F2025%2F03%2Fsomething_is_rotten_in_the_state_of_cupertino

I liked it. I'd say it was indirectly calling Melania a pig, but I don't think Trump fucks her. She seems to barely be able to stand him touching her. So, must be real pigs. I even suspect they're piglets.

Honest question. A lot of value comes from the StackOverflow answers. When a new technology comes out, and most people have migrated to LLMs for answers, where will the LLMs get new answers? If people are not discussing it on StackOverflow and elsewhere, would it just be docs? That could be problematic, since docs can be sketchy or non-existent. Will StackOverflow still exist in a lesser capacity to get answers on the absolute latest technology?

LOL! The funny part is you think it's obviously humorous. Um, no, we're talking Trumpers here. They believe all sorts of crap - Pizza Gate, QAnon ravings, humans don't cause climate problems, Trump makes a good president, and so on. But I agree having it be illegal is a bad idea - I'm just saying, it's some great propaganda for the lower half of the bell curve.

I thought so" meant I didn't expect to get any substantive, which turned out to be true.

I've been doing my own research, on my own, and more importantly, tasked by my employer, since the early LLMs. There's place where it can be somewhat useful, mostly for repetitive boilerplate type tasks, then there are claims being made, like yours, that I'd really like to see what's being done because it doesn't line up with my experience, and the experience of my cohorts. See my other post here for my opinion.

I could sit here and tell you all stupid crap I've gone through to try and get useful results from LLMs, but the end result is now, multiple times and at non-trivial expense, when the LLMs have been trained against our own codebase, it still doesn't create results that a junior dev can use effectively. So instead it creates half-wrong crap that I have to fix. Instead I could have written it myself, in freaking Notepad++, and it would have been more correct.

I don't have willful ignorance, I have disappointment and anger over over-hyped tech that fails to deliver. Will it get better, sure, but maybe? Or will this be the same stupid crap where a manager creates something in the LLM (like "code-free" tools of the past) that I end up either spending inordinate amounts of time fixing, or just do it myself.

'nough said.

I thought so.

Cool - show me something. Show me that happening, any way you want. Even show it to me failing admirably with it better on the next try.

...is he wants to push shitty software that costs less money to make. His job is to save money not make good software. Software is already pretty shitty, and at the point where it's produced in such a way that you try and fix bugs by trying to convince it with a conversation (prompt) it's going to get ugly. It's a very blunt tool for a very precise operation.

Some of us out here down in the trenches are actually trying to figure out how to use AI/LLMs to enhance software development. I'd certainly be willing to have him peek over my shoulder and see what it's actually like and what the actual results are and that the core concerns do not seem to be on the roadmap for future enhancements.

Honestly, so far all the AI/LLM stuff has been a productivity drag. It costs money in time, not to mention the energy costs, or just the fees associated with using it. Hard not to see it as a boondoggle.

Human's are making a great case against them being inhabitants of the Earth. Life will go on without us, and maybe the next sentient beings won't be so short-sighted, and they'll look back at our tiny blip on the geological timeline and shake their head(s) at what total idiots human's were, and not morn the loss.

And the poor people who buy into this chicanery, too.

Chasing rainbows can be satisfying in it's own right, I suppose.

...this would render the entire software development industry, and for that matter, all software, moot.

Want your own Google? Ask AlphaCode to make it. Same goes for Oracle Db, Twitter, any OS, anything! And your own AlphaCode? Ask AlphaCode to make it.

No one would be able to make any money from development of software, so the money will be in providing AlphaCode, until we have endless AlphaCode's all competing cause they make themselves. Then we're all just spamming endless amounts of AI generated software into the world.

Sounds great.

I don't think the terminology being used here is accurate, with sentient and brain being used, but, what if it was.

I can't help feel bad for this little under-powered being, trying its best to play pong but missing a lot of the time and just really struggling, because we couldn't make it any better with our current tech.

That number is way too low.