nico

10 ms·

181.
by nico 1y ago
Yes, when initially starting out the task/feature. But also most likely the agent won’t be able to fully do it on its own, and at that point the agent is not necessarily that much more useful than the chat Also, local agents miss conte
182.
by nico 1y ago
The article reads almost like an ad for o3 and spending a lot of money on LLM APIs In my experience, o4-mini-high is good enough, even just through the chat interface Cursor et al can be more comfy because they have access to the files dire
183.
by nico 1y ago
That's a clever solution I've been playing with chrome extensions recently, and have made them directly talk to a local server with a db. So using extensions, it's relatively easy to to store data locally and potentially syn
184.
by nico 1y ago
> Do we still need a back-end, now that Chrome supports the File System Access API on both desktop and mobile? Could this allow accessing a local db as well? Would love something that could allow an app to talk directly to a db that live
185.
by nico 1y ago
> it needs good tool calling, not an encyclopedic knowledge of the world I wonder if there are any groups/companies out there building something like this Would love to have models that only know 1 or 2 languages (eg. python + js),
186.
by nico 1y ago
> LLM in a sandbox using tools in a loop, you can brute force that problem Does this require using big models through their APIs and spending a lot of tokens? Or can this be done either with local models (probably very slow), or with sub
187.
by nico 1y ago
The stock market is a joke Yes, yes, I know it can stay irrational for longer than I can stay solvent But TSLA jumping almost 5% on news of failing earnings, is just ridiculous
188.
by nico 1y ago
> An early, basic version of a product (such as a piece of technology, a computer program, etc.) which meets the minimum necessary requirements for use by its creators and customers The keyword here is customers If you are building som
189.
by nico 1y ago
Super interesting, thank you for the link Aider writing its own code is definitely cool and within the same concept I’d love to see an LLM or some sort of coding model that modifies/trains the model itself
190.
by nico 1y ago
Great initiative I tried watching for a bit on makertube, but the stream kept dropping for me... it would work for 30-60s, then I would get a loading spinner. If I refreshed the page, and clicked the play button again, it would work ok agai
191.
by nico 1y ago
> Managing a codebase written by an LLM is difficult because you have not cognitively loaded the entire thing into your head as you do with code written yourself This happens with any sufficiently big/old codebase. We can never reme
192.
by nico 1y ago
For reference (according to Google): > The English Wikipedia, as of June 26, 2025, contains over 7 million articles and 63 million pages. The text content alone is approximately 156 GB, according to Wikipedia's statistics page. When
193.
by nico 1y ago
> If you know Factorio you know it’s all about building a factory that can produce itself This is a very interesting concept Could this be extended to the point of an LLM producing/improving itself? If not, what are the current limi
194.
by nico 1y ago
> That's a very simple example/context that I suspect most LLMs will be able to knock out with minimal frustration Yes an No. You are right that it's a relatively small project. However, I've had really bad experience
195.
by nico 1y ago
Very cool concept. Would love to see something like this for web development. Where making changes in the browser could be persisted seamlessly to code
196.
by nico 1y ago
It’s pretty good advice But the title is misleading. Sure, once you’ve built the habit of breaking bad habits, it will take 3 seconds each time. However, it will take quite a bit to build that habit The article references Dale Carnegie. Rel
197.
by nico 1y ago
I don't know about the photo-management aspects. However, I've had very good experiences running gemma3 (4b and 12b) locally via ollama I've used gemma to process pictures and get descriptions and also to respond questions ab
198.
Meet Gen-Xi – By Jason Calacanis(theallinpod.substack.com)
2 points by nico 1y ago|0 comments
199.
by nico 1y ago
How big would the telescope/mirror/lens need to be to get a picture of something in the Alpha Centauri system, 4.37 light years away? Also, could the image be created by “scanning” a big area and then composing the image from a bu
200.
by nico 1y ago
Humans work pretty much the same way Several studies have shown that we first make the decision and then we reason about it to justify it In that sense, we are not much more rational than an LLM
201.
by nico 1y ago
> 2.3 Specialized Micro-Agents Over Generalized Rules Initially, our instinct was to continuously add more rules into a single large prompt to handle edge cases This has been my experience as well. However, it seems like the platforms li
202.
by nico 1y ago
Immutable functions! I guess that’s one way of doing functional programming /s
203.
by nico 1y ago
This is key Besides model capabilities, one of the most important aspects of AI-assisted development right now, is context management Cursor et al try to automate that for the user, and it works up to a point. But at a certain level of comp
204.
by nico 1y ago
Does it need a newer GPU? Or can it run on just CPU? Would it run on a raspberry pi?
205.
by nico 1y ago
This is a really cool feature and it’s big competition for services like Lovable, Bolt, v0 Seems like AI-assisted coding space is splitting in 2: 1) tools and services that aim mostly at prototyping and are close to no-code; most useful for
206.
by nico 1y ago
> Proof-of-work CAPTCHAs work well for making bots expensive to run at scale “Expensive” depends on the value of what you do behind the captcha There are human-solving captcha services that charge USD 1 for 1k captchas solved (0.1 cents
207.
by nico 1y ago
After clicking the button "View full video" I got: "Could not play video. Please ensure you have interacted with the site first and try again." PS: the cybertruck in the video looks a lot better than a real one
208.
by nico 1y ago
Great concept and very relevant today[1] It’s interesting that using LLMs is making very explicit that “someone” needs to read the code and understand it. So having good comments and making code readable is great both for AI and humans 1: “
209.
by nico 1y ago
Thank you YC for posting this before the talk became deprecated[1] 1: https://x.com/karpathy/status/1935077692258558443
210.
by nico 1y ago
For reference, Terminator 2 is from 1991

More ›