Honestly it has become less and less useful over the past 15 months. I imagine it’s due to the EU AI Act, but it could just as well be a failure in leadership.
It repeats itself continuously, even when specifically asked not to. And it just doesn’t come up with interesting stuff any more. It always uses the numbered list format even when I specifically ask it not to. The custom ‘app’ gpts are a mess and have only wasted my time. I consider canceling my subscription regularly 😕
I should add that I’ve had early access and built stuff on top of it and it was extremely exciting. Now I am just curious about future foss solutions to get back to experimenting for real 🌠😁
For now I’ll likely just switch to another product, as it’s definitely useful to get an overview when learning. But to be frank a good book is just better 🤔
it’s not like things like this haven’t been posted consistently since chatgpt started on the scene
oh no, wait, I meant the other thing. the one where it’s always been this awful dogshit.
“imagine it’s due to eu law” what the FUCK? I’ve seen some wild takes but wow, you’ve just strength-hammered the shit out of the fair bell in a way I haven’t seen in a while
Well over a year ago it was actually useful. […] And it just doesn’t come up with interesting stuff any more.
i have to admit i’m deeply curious what outputs you considering interesting enough for twenty bucks a month
love how they edited their comment to be “more specific” and just made it even worse
“It’s bad because of EU regulation” is the new “it’s bad because it’s woke”.
@gerikson @techtakes I think you’ll find Boris Johnson pioneered that one in the early 1990s.
Was that before or after hairdressers gave up on him?
True, but it’s recently crossed the pond to Silicon Valley. I think it was when the DMA affected Apple that a lot of hackernews became EU regulation experts and started grappling with the fact that laissez-faire is seen as a dirty word in the country where it originated.
I actually appreciate the EU regulations. The real question is whether companies are creative enough to come up with new solutions that fit.
Do any “ai” companies have a business plan more sophisticated than
- steal everything on the web
- buy masses of compute with vc money
- become too important to be busted for mass copyright infringement
- ?
- profit
I don’t recall seeing any signs of creativity, or even any good ideas as to what their product is even for, so I wouldn’t hold my breath waiting for one of the current crop to manifest creativity now.
Perhaps I missed something, though?
I get your point but I feel it goes a tad bit too far. It’s like that now, but it used to have the capability to adapt well and also come up with original ways of combining things. I actively leveraged that capability for my library. Rn it’s unfortunately on pause due to the regression, at least within the NL/EU.
you appear to be lost, this isn’t “sama stan club”
and by library I hope you don’t mean a place of knowledge other humans rely on
you’re learning from an autocomplete
Well, I had three electronics experts who validated my design based entirely off what I’ve learned from gpt4.
you just really don’t get it, do you?
thanks to the EU, those three electronics experts are now 1.5 door-to-door vacuum cleaner salespeople
I’m satisfied with the design and the sparring we’ve done.
- attempt to learn from llm
- have llm tell you about success criteria
- get some or other shit done
- ???
- be personally satisfied, individually
new electronics design standard everyone! no more need for EDA sims, we can just ask this clever poster for verification! progress! i’m sure it’ll be perfectly fine!