@kentcdodds@mattpocockuk Disagree on Kent's take at least for now if you have seen some the React code agents output you should have some understanding of react to know it's a bad practice and your codebase will collapse if repeated
@QuintenFrancois The worst part is either he is very ignorant and a complete imbicile or he knows perfectly well what he's doing. And I'm afraid it's the latter. Fucks up the country just for the game.
@mjovanovictech@johncrickett@Revolut The openweight models are good enough for most tasks they will always lack a bit behind on the frontier models. We did the same at my company.
@DavidSaranga Ideally the teacher should be neutral and teach them about the genocide Israel is committing, but also about the terror attacks of Hamas.
@plainionist I don't think there is something like true creativity everything is always inspired by something else. And don't get me wrong I'm not downplaying creativity, but that's essentially what it comes down to.
@ericweinstein Taking a lesson from Israel, ok so we just murder them all then? They are being called genocidal because they ARE. That being said I agree the eu has been far too lenient to immigrants and this needs to changes ASAP
@nico_vrgwn@Mauritsvdr Welk een randdebiel ben jij. Ik heb niet eens iets over Tesla gezegd ik praat over de US in het algemeen. Ik werk voor een Benelux bedrijf, maar werk vaak samen met bedrijven uit de US. Het feit dat je zelf voor een bedrijf uit SF werkt zegt al genoeg.
@Xinirnix@pbl999@Mauritsvdr Inderdaad geen idee waar die mensen het halen. Maar de bio van deze persoon zegt al genoeg, moet dringend eens uit zijn of haar echo kamer komen.
@nico_vrgwn@Mauritsvdr Ik weet niet onder welke steen jij leeft, maar Maurits heeft wel degelijk gelijk. Innovatie gaat veel sneller in de US. Het is niet dat er niets innovatief uit de eu komt, maar we hinkelen wel achterop.
Or stop being overly reliant on your LLM and make sure you have a simple workflow that works reliably no matter what you're using.
They also have 0 incentive to optimize your token spending.
a weird inversion with LLMs is the models improve faster than the tinkerers
when i see people with custom workflows and setups they're all addressing problems that don't exist anymore
the person naively using vanilla codex is more likely to be experiencing state of the art
@DanielGlejzner Yes definitely, it's way more reliable. consistent results beat "state of the art". They can't even tell what state of the art means because they have no way of checking.
@mattpocockuk@atmoio Either @atmoio has never worked in a team or doesn't understand what skills are. @mattpocockuk and @poteto skills work great in teams and sure they are not going to magically fix everything you still need to know what you are doing. Having great results so far.