nico

19 ms·

331.
by nico 1y ago
The video linked in the post description shows everything pretty clearly, here: https://www.youtube.com/watch?v=JlCktXTY8sE
332.
by nico 1y ago
Great use case. And very funny situation with the reasoning models! :)
333.
by nico 1y ago
Amazing. Thank you for the great resources!
334.
by nico 1y ago
Looks amazing. Congrats on the release How does this compare to browser mcp ( https://browsermcp.io/ )?
335.
by nico 1y ago
Thank you for the numbers What have you used those models for, and how would you rate them in those tasks?
336.
by nico 1y ago
Great recommendation about the community Any other resources like that you could share? Also, what kind of models do you run with mlx and what do you use them for? Lately I’ve been pretty happy with gemma3:12b for a wide range of things (ge
337.
by nico 1y ago
How long can the recordings be? Recently needed to do some 1-2hr long recordings. Look works pretty well for that, QuickTime + OBS could, but generates massive files, which are hard to share
338.
by nico 1y ago
Compilers, for the most part, made it unnecessary for programmers to check the assembly code. There are still compiler programmers that do need to deal with that, but most programmers get to benefit from just trusting that the compilers, an
339.
by nico 1y ago
To people curious or skeptical if this could be called “seeing” or “hearing”, I recommend listening to the Batman podcast episode on NPR ( https://www.npr.org/2015/01/23/379134306/batman-pt-1 ) Through the
340.
by nico 2y ago
The qwq-2.5-0.5b is the tiniest useful model I've used, and pretty easy to fine-tune locally on a Mac. Haven't tried it on an iPhone, but given it runs at about 150-200 tokens/second on a Mac, I'm kinda doubtful it could
341.
by nico 2y ago
Would love to see it doing some work on a Google spreadsheet (including doing formulas, vlookups, data import and cleanup) and then creating a decent Slides presentation with some charts from the spreadsheet
342.
by nico 2y ago
Thank you, will try out your suggestions Have you used something like a director model to supervise the output? If so, could you comment on the effectiveness of it and potentially any tips?
343.
by nico 2y ago
I agree with you. The outputs are usually good, it’s just that for the use case I have now (writing several pages of long dialogs), the output is not as long as I’d want it, and definitely not as long as it’s supposedly capable of doing
344.
by nico 2y ago
Thank you for the insights and useful links Will keep experimenting, will also try mistral3.1 edit: just tried mistral3.1 and the quality of the output is very good, at least compared to the other models I tried (llama2:7b-chat, llama2:late
345.
by nico 2y ago
Been playing with that, but doesn’t seem to have much effect. It works very well to limit output to smaller bits, like setting it to 100-200. But above 2-4k the output seems to never get longer than about 1 page Might try using the models w
346.
by nico 2y ago
They are multimodal. Havent tried the QAT one yet. But the gemma3s released a few weeks ago are pretty good at processing images and telling you details about what’s in them
347.
by nico 2y ago
What kind of functionality do you need from the model? For basic conversation and RAG, you can use tinyllama or qwen-2.5-0.5b, both of which run on a raspberry pi at around 5-20 tokens per second
348.
by nico 2y ago
Been super impressed with local models on mac. Love that the gemma models have 128k token context input size. However, outputs are usually pretty short Any tips on generating long output? Like multiple pages of a document, a story, a play o
349.
by nico 2y ago
Thank you for the insightful reply > There's no equivalent to "does everything kinda well" like chatgpt or Gemini on local, except maybe the 70B and larger, but those are slow Is there something like a “prompt router”, tha
350.
by nico 2y ago
Any specific model recommendations for running locally? Also, what tasks are you using them for?
351.
by nico 2y ago
Or maybe figure out a different architecture Either way, the end user experience would be vastly improved
352.
by nico 2y ago
> For me, a team of junior developers that refuse to learn from their mistakes is the fuel of nightmares. I'm stuck in a loop where every day I need to explain to a new hire why they made the exact same This is a huge opportunity, m
353.
by nico 2y ago
> It's... decidedly expensive to run an LLM this way right now Does it work ok with local models? Something like the quantized deepseeks, gemma3 or llamas?
354.
by nico 2y ago
Not sure about denaturalization, but this is getting pretty extreme already (USCIS, effective immediately, is screening immigrants for criticism of Israel): https://www.reddit.com/r/UnitedNations/comments/1jw8
355.
by nico 2y ago
> fraudulently obtaining citizenship It seems like they are claiming that “not saying you intended to protest” is essentially lying on the initial visa application, thus fraud and grounds for revocation of residence and deportation So, i
356.
by nico 2y ago
> might as well assert your rights I wonder what could happen to citizens who weren’t born in the US Given what we’ve seen recently. Could it be possible they would refer these people to the state department to revoke their citizenships?
357.
by nico 2y ago
Just did a test last week and OpenAIs research was way better. Found 10x more sources and did an overall pretty great job The task was to lookup information about a late distant family member who had been a prominent employee in a certain f
358.
by nico 2y ago
> This will probably mean you have some revenue and some sense of where PMF might lay or some kind of brand/pedigree. > Do not underestimate the value of YC brand and being able to present on Demo Day gives you Except to get into
359.
by nico 2y ago
My personal experience with pair programming is that I first was very intimidated by it and instinctively hated it Then, it slowly creeped up at work, and I noticed that initially, most people seemed uncomfortable with it. But, if we were a
360.
by nico 2y ago
As mentioned in other comments, this is just a concept, they haven’t built it But it’s a pretty cool concept. Video here: https://m.youtube.com/watch?v=Q8TUTp0_D-I

More ›