4 ms·
51 comments so far, the vast majority panning Astra's coding abilities. An uninformed reader may come away with the impression that this isn't an absolutely rev
by nvrmnd 22d ago
51 comments so far, the vast majority panning Astra's coding abilities. An uninformed reader may come away with the impression that this isn't an absolutely revolutionary technology that with coding abilities many of us thought were not even going to be possible with language models as recently as a year ago.
Yeah, it's not perfect, but it's really good and extrapolating this rate of improvement for 6 months is rather terrifying (from a SWE perspective, at least).
- sbt 22d agoThere was a big jump around new year, but they seem to have flatlined since them. Just my experience.
- km144 22d agoThe biggest jump was Opus 4.6. Since then they have gradually gotten better at finding issues in your reasoning, not hallucinating, and being rigorous with the code, but much much worse at explaining things and generally just talking in a way that a human can understand. All the models I've tried seem to be suffering from the same fate so it must be something going on with the training meta right now.
- applfanboysbgon 22d agoIt's impressive technology but not revolutionary. Revolutionary technology would have resulted in, you know, a revolution in software quality. Instead quality keeps going down.
- oleggromov 22d agoIt is revolutionary. Programmers are paid less to work more.
- byzantinegene 22d agoi think the main consensus here is that the actual performance is not indicative of the benchmark performance (which supposedly outperforms the previous iterations)