strange_quark

After the past week or so, I wouldn't be surprised if we end up with professional licensing for software engineering.

The thing I find least compelling about these “AI is a compiler” arguments is that they require dissonance or magical thinking where somehow the AI is good enough to write and review all the code to the point where you don’t even need to look at it, but at the same time the LLM needs an experienced software engineer to feed it the prompts.

Idk, I see it the opposite, while they can sometimes write impressive-looking code under the right circumstances, I still see the most advanced models screwing up the basics all the time. It seems to me there is some fundamental flaw in LLMs that is simply not fixable at any cost. While we can keep pushing what they can do (or rather what we allow them to do), 1. English just isn’t a good programming language and 2. There’s no there there.

Yes, they're very good at writing code now. The problem is that we (or at least I am) drowning in code that runs and does something that resembles the problem you asked it to solve, but that something is frequently semantically incorrect.

Something has to give; the constant onslaught of slop and workflows changing every few months just isn't sustainable. Either we slow the fuck down and figure out what we're even doing here, or the AI improves so much that we're all out of a job.

No. TFA says one was from October 6. You know, after they all got so scared and decided to “pace the frontier”.

> What employer would be willing to pay for so much wasted time for their software engineers to do deliberate training courses, or willing to accept the productivity loss from their engineers occasionally working manually for practice, when working manually is so much slower?

Lots of regulation or some landmark lawsuits should do it. If airlines operated the way tech companies did, we'd have daily plane crashes.

Pilots have to know all the ins and outs of every single system and subsystem of the machine they operate. If software engineers were required to understand their systems to the same degree in order to ship something, the only way you could do it is by coding yourself.

Dario’s wife literally begged Epstein for money to do a luxury porn site for women.

In fact, a lot of AI research was funded by Epstein!

These apps have to be astroturfed, right? Like I get the desire to give Microsoft and Adobe the middle finger and I guess it's somewhat impressive these apps even do anything, but man are they such buggy pieces of trash. Like you said, I downloaded the Word clone and it's broken in so many ways that I can't even describe what the bugs are. In 5 minutes of messing around, I hit so many random and inconsistent failures of basic functionality like styles, image resizing, and table formatting. Plus the UI is atrocious. All the downsides of Windows apps mixed with all the downsides of a weird cross-platform UI framework.

How much time do we have to wait before we are allowed to criticize AI? GitHub copilot has existed for half a decade. Cursor, Claude code, codex, have all existed for over a year and have seen more investment that practically any product in history and have almost 0 barrier to entry. Where is my decade of software progress in 1 year, because that’s what 10x productivity looks like! All I see are the same old broken products and services from basically the same handful of big players, higher costs, and more bugs. Remember when every year iOS and Android saw loads of new features and a radical rethink every few years? And that happened on ancient hardware and in a significantly smaller industry. Where is anything like that happening now?

Matches my experience to a tee.

And don’t forget the company also spent a bunch of money in tokens for the initial author to implement the thing poorly.

I don’t think that’s sufficient. If they want to be good stewards of mathematical research and not just doing marketing, they need to at the very least tell us the datasets and any techniques they used to train this model. IMO that’s probably just as if not more valuable than a dump of un-reviewed results.

They won’t because they don’t care and the only way it got this good is something along the lines of they trained on every mathematician’s codex sessions even if they opted out because they consider the thinking traces or output and metadata fair game.

Too conspiratorial to think that they (the AI companies) want to be blocked or at least temporarily delayed from IPO? Finances are too bad to do it but delaying for ambiguous reasons will sink their valuations. Trump can’t let the stock market collapse before the midterms. Hence this whole big safety dog and pony show. Keep the market up until the midterms, then hand off the bag.

$DAYJOB recently introduced a AI writing policy because people were sending each other mountains of slop back and forth enough that it became a huge time suck. The policy is basically: don't, with the justification being "writing is thinking". It's like they're so close to getting it.