chenxi9649

6 ms·

31.
by chenxi9649 3y ago
I wonder if it also had to do with the fact that handwriting is slower . Which means that you generally hold the concept in your working memory for longer than typing. Which often times does not require the concept to be held in working me
32.
by chenxi9649 3y ago
Excited to see this on HN front page! Unlike others, I didn't mind the CC + 30 day trial. 30 days actually feels generously long as I don't see many other projects offering that. I think it's ultimately right move for this pr
33.
by chenxi9649 3y ago
Love what you guys are doing. Is it possible to host custom models on fal?
34.
by chenxi9649 3y ago
When I don't wear contacts for a while(just normal glasses), wearing them throughout the day also makes me a bit "tired". Even though the "resolution" is basically the same. But after wearing them consecutively for
35.
by chenxi9649 3y ago
depends on what "in the end" means haha
36.
by chenxi9649 3y ago
I guess a good enough AI agent is essentially a human worker. I wonder if all the work that's being put in right now by agent projects will become more or less "useless" similar to those specialized classification models be
37.
by chenxi9649 3y ago
Yea that's pretty cool too, I heard some restaurants are also doing a 100% voice LLM to take orders. Transcription, specifically Whisper, is one of those ML models where the accuracy is basically on-par with humans. So, I really expect
38.
by chenxi9649 3y ago
Good point, I should've defined this a bit more clearly in the post. Honestly, I'm not toooo sure how to segment the term "agents", but in my mind there seems to be one realm for retrieval assistance. Ie. how do we make
39.
by chenxi9649 3y ago
This is pretty cool! Another use case where the cost of being slightly worse than a human is totally fine.(coming from someone that doesn't write tests lol) I'd love to learn in more detail how it created those factories, correcte
40.
by chenxi9649 3y ago
Looping back to what the other person was talking about -> "Areas where slightly lower accuracy is acceptable." Seems like information retrieval of any sorts is one use case where the cost of being wrong is not super high. I gu
41.
by chenxi9649 3y ago
Interesting, it seems that using an LLM as an agent to help with knowledge retrieval is one concrete use case that I've seen people do repeatedly. It also feels like we are at a bottle neck when it comes to the knowledge retrieval prob
42.
by chenxi9649 3y ago
>even the latest GPT-4/Gemini doesn't produce anything I can drop in and have it work This is certainly true for more complex code generation. But there are a lot of "rote" work that I do use GPT to generate, and I fe
43.
by chenxi9649 3y ago
curious to know, which tools do you use and how do you use em?
44.
by chenxi9649 3y ago
Yea that's a good point. Now I am curious, what are some tasks that can accept a model that is at 80% as good as a human, but is 100x cheaper?(or, 100x faster?)
45.
Ask HN: What are some actual use cases of AI Agents right now?
169 points by chenxi9649 3y ago|149 comments
46.
by chenxi9649 3y ago
Also saw this on Twitter a few days ago https://www.pinecards.app/ Haven't tried it myself but it seems like a more modern/gamified version of Anki. I wish there was a market place for spaced repetition decks on d
47.
by chenxi9649 3y ago
Just because we are all gonna die, doesn't mean we shouldn't try to die slower... Or die more happily. Just because you're going to poop doesn't mean you shouldn't eat. HN is a tech forum yes, but it's also peo
48.
Ask HN: What's the state of LLM knowledge retrieval in Jan 2024?
20 points by chenxi9649 3y ago|2 comments
49.
by chenxi9649 3y ago
It's still not too clear to me when we should fine tune versus RAG. In the past, I used to believe that finetuning is mostly for model behavioral change, but recently it seems that certain companies are also using fine-tuning for knowl
50.
by chenxi9649 3y ago
Reallllllyyyyyyy..... checking it out now
51.
by chenxi9649 3y ago
It's crazy to me how well this works, especially on mobile. If league or even TFT was as accessible as a tab like this it'd be the end of me.
52.
by chenxi9649 3y ago
I believe one thing they're planning on doing is selling the CO2 to certain concrete manufactures. Another thing is putting the CO2 back underground through certain processes!
53.
by chenxi9649 3y ago
Are there any benchmarks/results showcasing the performance increase with fine tuning? Like showcasing a list of prompt and answers between a fine-tuned LLM and a non fine-tuned one. Could be with LLAMA 2 as well. Even if it's qui
54.
by chenxi9649 3y ago
Looks pretty good!Though there seems to be a new tailwind ui libraries that looks good every month. What's the differentiating factor between this and shadcn/nextui/daisyui/tailwindui/flowbite just to list a few on
55.
by chenxi9649 4y ago
Normal/average talking is ~150 WPM. Average typing speed is about 60-70. Is a 150+WPM a requirement to become anonzzies' friend?
56.
by chenxi9649 4y ago
This is very well made and designed. I will likely use this instead of the actual Chatgpt UI since their API is a lot cheaper than the 20$/month pricing for my usage amount. Interesting note: I tried speaking mandrain chinese to the mi
57.
by chenxi9649 4y ago
I don't think you're the only one who's experienced this, but I think you're in the minority in realizing that you're experiencing this. This thread has a lot of gold nuggets, but most seem to be focused on how to s
58.
by chenxi9649 4y ago
I originally thought it was something like DaisyUI but for React Native with NativeWind, but this seems quite different, and almost a little strange haha. I'm very impressed by the polish of the project but am curious about the main us
59.
by chenxi9649 4y ago
Same for us. Certain dev environments bugs were been a big pain in the ass for us in the past. For the logic side of thing React Native has been a nice experience for us as well!
60.
by chenxi9649 4y ago
I think Elon lovers might say that his experience with Paypal/Zip2 and more recently Starlink gives him decent credibility regarding software infra and networking. However, being up on top and writing the code are very different things

More ›