yamajun93

3 ms·

1.
by yamajun93 5mo ago
I don't think the problem itself is being AI-generated, but lazy AI-generated UI/UX often is. The vibe-coded aesthetic often signals "I didn't deeply consider how this actually works" - which can be a red flag even
2.
by yamajun93 5mo ago
This is so cool! I wonder how you keep the context so it remembers what it was rendered before. Since it is real-time, how do you keep consistency in the webpage so user can browse back and revisit already rendered sections?
3.
by yamajun93 5mo ago
This is a real pain, I notice over-editing a lot when using AI coding tools locally and also re-inventing the wheel (e.g not using external libraries or reimplementing something over and over again). The Boy Scouting Rule is a good framing,
4.
by yamajun93 5mo ago
How come pseudoanonymous does not exist? There are definitely some data that can trace the user without factual data https://www.ru.nl/en/staff/researchers/research-data/persona... And well, if I'm
5.
by yamajun93 5mo ago
The distinction between anonymous and pseudoanonymous matters a lot in practice, even with a unique machine ID hash, there are patterns can still identify users across sessions. Would be good to see a clearer breakdown of whats collected an
6.
by yamajun93 5mo ago
I wonder how the focus of agentic-ai differs from that of "normal" LLMs calls in terms of hardware. Does this just provide faster TPU, or does it support it in other ways?
7.
by yamajun93 5mo ago
The benchmarks looks great for a 27B model, but Im curious how it will perform locally. I have tried a bunch of open-source models, and still I feel we are far from getting a similar output to claude code running in my M3 in my macbook pro.