201–210 of 442 posts
I'm not entirely sure why AI knowledge must be close to a year old, and clearly this is a problem developers are aware of.
Is there are a technical reason they can't be, for instance, a month behind rather than close to a year?
Earlier quoted context omitted.
After watching the entire world’s reaction to AI, at this point my conclusion is that hype driven development is human nature, and we just need to come to terms with that (but you will have to drag me kicking and screaming).
Maybe if you think artificially inflating the hype through massive ad campaigns, marketing campaigns, and shoe-horning AI into every product then yeah the world has a reaction to AI. It's mostly been meh, things like Apple Intelligence and Office Copilot have largely fallen flat. If the hype was real none of these AI initiatives would be struggling to make money, but they are. I don't really see it different than the…
It's also little different than the .com bubble...
I think this teaches us that a thing can be hyped into the stratosphere, suck very much, crash and burn, and then go on to eat the world...
I noticed this as I experimented with alternatives for React and all of them I tried were terrible on OpenAI/ChatGPT. Either it doesn't know them, or it makes weird mistakes, or uses very outdated (no longer working) versions of the code. It is also annoying that most modern JS things have 4 versions to do the same thing: With TS, With TS + Decorators, With plain JS, with JSX, etc. so code generation picks one that i…
It just makes up Cmdlets a lot of the time. If you prod it enough though it will eventually get it right, which strikes me as odd, it's like the training data was just full of really bad code.
By contrast, anything I've asked it to do in Python has been more or less spot on.
I fear that in the future the choice of tech stack is going to be less on the merits of the stack itself and more centered around "Which language and framework does ChatGPT (or other AI) produce the best output for"
This is truly terrible. What happened to a new JS front end library every week? If this keeps up, we won't get to completely throw away all of our old code and retool every two years (the way we've been operating for the last 20 years) How will we ever spend 85% of our time spinning up on new js front end libraries? And don't even get me started on the back end. If AI had been around in 2010, we probably still have s…
As if web/css wasn't a new gui library no one asked for. Peak programming was VB6 & Delphi 6 (maybe 7). Everything after that was just treading water in increasingly degenerate ways.
know what this will be about without reading Python 3.12-style type annnotations are a good example imo, no one uses the type statement because dataset inertia
Type annotations by themselves, are little more than a comment
Also, each package should ideally provide an LLM ingestible document. Upload this for the LLM, and have it answer questions specific to the new package.
This is truly terrible. What happened to a new JS front end library every week? If this keeps up, we won't get to completely throw away all of our old code and retool every two years (the way we've been operating for the last 20 years) How will we ever spend 85% of our time spinning up on new js front end libraries? And don't even get me started on the back end. If AI had been around in 2010, we probably still have s…
Earlier quoted context omitted.
> And I also found the basic tier of ChatGPT hallucinated my app's purpose in a way that was borderline slanderous. I'm curious about this, what exactly did ChatGPT write and how was it borderline slanderous? Sounds like a big danger.
So ChatGPT seemingly guessed its purpose just from its name. Its name is CurrentKey Stats, and it inaccurately described it as an app that kept stats on the current keyboard keys you were pressing, so essentially a key logger, which is again completely wrong. I was actually somewhat hesitant to make this comment out of the fear that the next AI models will train on it and reinforce the false idea that that's what my…
It's not like the bad old days where sentiment analysis was a bag of words model, add up all the "positive" words and subtract from that total the number of "negative" words — back then, they would mis-identify e.g. "Product was not as described, it did not come with batteries, and the surface wasn't even painted" as "this review favours this product" because they couldn't handle "not" or "wasn't".
Earlier quoted context omitted.
Not if they're trained to work through the problem each time they encounter it and stay up with their clinical training. The day the new classification drops many have already heard about it. You also assume that all the models in use will in fact be retrained. Generally, this position flies in the face of lived experience. AI is in fact stifling adoption of new things across many industries.
There's no way that doctors in the US continue their training. I suffer from a chronic illness. I saw 5 different specialists over the past 2 years, and each one gave me different treatment. A couple even relied on information from the 70s and 80s. One even put me in the ER because he changed my treatment after I explicitly told him not to. Another example: Back in my 20s, I injured my back. I took my MRI results to…
What about closed source tooling? How do you expect an AI to ever help you with something it doesn't have a license to know about? Not everything in the world can be anonymously scraped into the yearly revision.
If AI is going to stay we'll have to solve the problem of knowledge segmentation. If we solve that, keeping it up to date shouldn't be too bad.