edg5000

6 ms·

91.
by edg5000 3mo ago
Great speeds. So a small model on a lot of slow ram is fast. Now I wonder how larger models run on that thing.
92.
by edg5000 3mo ago
I do about 10 google search queries for every 1 opus/gpt prompt. For google, I don't actually open pages anymore 9 out 10 times; I rely on the AI summary. It's fast and accurate; the trick is that you learn where the boundary
93.
by edg5000 3mo ago
I ask it to propose revisions, which I then review and provide detailed feedback on. It then works, and afterwards I review. I always commit beforehand. This works well for me.
94.
by edg5000 3mo ago
I've experimented with completely novel tool call syntaxes combined with ordering and structuring the context differenly, and they seem to have a high tolerance for this stuff. But you always risk losing performance, probably highly un
95.
by edg5000 3mo ago
I spent 6 months frantically working fulltime on a harness, and when it was finished, I just stopped using it and went back to Codex CLI. I've also stopped using all the fancy features: subagents, planning mode. I've also stopped
96.
by edg5000 3mo ago
Their PCB service must be a loss leader, but their assembly service might be profitable. What's also a loss leader is their default configuration. Want .2mm vias? That's another 30$. Want to assemble both sides? More fees. So it g
97.
by edg5000 3mo ago
I like how they're using real numbers and theorizing around that. An interesting next step would to dive in deeper, to see up close where it's bottlenecking.
98.
Chipmakers have 67,000 unfilled jobs. An old training trick won't work this time(qz.com)
3 points by edg5000 3mo ago|0 comments
99.
by edg5000 3mo ago
Even if it's uploaded once, it's still being ran through inference. It saves a bit of HTTP traffic I suppose.
100.
by edg5000 3mo ago
None of them upload the whole repo, which is what this link claims it does. That's unheard of.
101.
by edg5000 3mo ago
Wow, very interesting fact. If a unit is performing badly due to lack of equipment, they won't get the points to get the equipment they need. I wonder if that's a problem in practice. I'll read up on it, very interesting stuf
102.
by edg5000 3mo ago
Fun sidenote: On a recent Hotmail post on here, I saw a screenshot of hotmail V1 (way before M$ took over). They had a prominent button on their homepage about privacy.
103.
by edg5000 3mo ago
It scores high on BenchCAD, that's interesting to see, I was wondering about how each model could handle this. Seems like they trained it on programmatic CAD specifically.
104.
by edg5000 3mo ago
This comment is odd, not conventional — it reads suspiciously verbose, not consice.
105.
by edg5000 3mo ago
We can look at openrouter GLM 5.2 prices I think to get a rough idea about the pure inference cost (with margin). They are probably running on simmilar hardware. Although scale helps Anthropic probably.
106.
by edg5000 3mo ago
Because it'll bottleneck at the CPU? Or because 200 bB/s is overkill when running a model (that fits in memory)?
107.
by edg5000 3mo ago
Those font tests are cool, he's created a unique way of rendering fonts, an interesting, playful look. And great that he's letting his son play with the computer.
108.
by edg5000 3mo ago
Any good LLM will emit Python or other scripts to analyze or work with large files (e.g. GeoJSON) naturally when asked to work with a large file. So I don't see that as needing an explicit solution. The LLMs just figure it out.
109.
by edg5000 3mo ago
Your argument is that it's simply because of dropping prices. Maybe that does explain the bulk of it. But the article seems to suggest it's the WFH transition. How much of the vacancy rate would be explained by that?
110.
by edg5000 3mo ago
For me, the encrypted reasoning contents, when looking at the base64 string lengtht, show this effect. However, the server-reported reasoning tokens don't. So I assumed it was part of the encryption and/or obfuscation purely. So I
111.
by edg5000 3mo ago
There seems to be no awareness from EU govenments about how much power we're handing over to two large outside companies. This incompetence in the leadership will cause a lot of harm over the years. This has been going on for a long ti
112.
by edg5000 3mo ago
That he ran into blockers with the exotic setup is understandable. After overcoming a bunch of challenges, I was surprised to read that the missing Flatpak package stopped him in his tracks, that felt surprising when I read it. It feels lik
113.
by edg5000 3mo ago
> there was no org.freedesktop.Platform.GL.nvidia in Flatpak repositories for AArch64 All he had to do was build some packages from source, right? It's really worth learning how to do this, since it removes a lot of constraints. And
114.
by edg5000 3mo ago
You also need setuid, that's were I hit the root requirement, the bind mount itself can indeed be created.
115.
by edg5000 4mo ago
Bind mounts can work fine. Setting them up does require root though. Easiest would be if the harness offered to enable containment. Awkwardly, it would require root.
116.
by edg5000 4mo ago
> several barns > 1/20th of a human brain It's very impressiive how accurate his predication is. Especially on the training part.
117.
by edg5000 4mo ago
It's in the JSONs in ~/.claude, but last 30 days only I think. You can have the model analyze history. So for correct history you'd need to run history analysis on a cron job or something. Kinda hacky.
118.
by edg5000 4mo ago
By stating you used an LLM it sounds as if you haven't vetted it. By mentioning the LLM in this hyped-up time, you're distracting from the actual deep work you've (presumable) done. By ensuring the patch is truly tight, and t
119.
by edg5000 4mo ago
I see. So not quite usable apart for specific use cases. Maybe in a few years we'll see new hardware players and better prices.
120.
by edg5000 4mo ago
One advantage about local LLM: You could serialize the context yourself, without being constrained by APIs. And let's not forget, the Big 2 encrypt their thinking. If you use custom clients, which is a very grey area alreay, being able

More ›