edg5000

12 ms·

211.
by edg5000 6mo ago
as a starting point, asianometry has some good videos on this
212.
by edg5000 6mo ago
Yes, it's a recurring pattern also seen in European politics.
213.
by edg5000 6mo ago
I always felt this moment would come eventually. The trend is centralisation of power and control. It's depressing. It's been a long time coming at a slow but consistent cadence.
214.
by edg5000 6mo ago
It's always the same pattern. There is no way to protect the children while also preserving freedom. The rationale behind it is irrelevant. For this to work everything would have to be locked down right? This is not in the interest of
215.
by edg5000 6mo ago
Using a coding CLI I was able to create a computed shape would otherwise by incredibly challenging to make (impossible in parametric CAD GUIs at least) LLMs + CadQuery is a powerful combo, this will be more and more common I think. It'
216.
by edg5000 6mo ago
What can and what can't it do compared to Codex and CC?
217.
by edg5000 6mo ago
The main issue was that they forked UBoot and did not release their modifications, making it hard to run anything other than their Armbian fork. They forked Armbian a long time ago and kind of hacked things together rather than adding suppo
218.
by edg5000 6mo ago
Power is taken but also given. It's a dynamic and I agree it's gotten way, way out of hand. It may eventually supress progress and become a real parasitic presence, but we've not reached that point yet (in net terms). Google
219.
by edg5000 6mo ago
If the plugins were bought for six figures, then it must be incredibly lucrative. How on earth could they be making it back? Is injecting spam into Google results THAT lucrative?
220.
by edg5000 6mo ago
> In 2017, a buyer using the alias “Daley Tias” purchased the Display Widgets plugin (200,000 installs) for $15,000 and injected payday loan spam. Is that it? Going through all that trouble just for some spam? Surely more lucrative crimi
221.
by edg5000 6mo ago
Cool use case. I have it PDFs for my electronic load and lab PSU, and it was able to write drivers in C++. Quite a powerful use case. I did battery discharge tests this way.
222.
by edg5000 6mo ago
For large georeferences textured meshes, generally 3D Tiles is used which uses a hierarchy of oriented bounding boxes to support various LODs. If you split up your model into chunks as well as diferent LOD levels, the viewer can then reques
223.
by edg5000 6mo ago
They do a have a vision decoder like many other LLMs, so in theory it should be able to write the positions textually, then call a render command, then look at the rendered bitmap. I's all very opaque though; I'd love a visualisat
224.
by edg5000 6mo ago
They all have Intel or AMD chips. My surprise is that there are no good, non-US ARM options. I'd love to be proven wrong though.
225.
by edg5000 6mo ago
Interestingly, there are zero non-US powerful laptops. The closest option is the Moore Threads MTT AI Book (12-core 2.65Ghz, 32GB DDR5, 1TB SSD, 14 inch). It cannot reach a modern Ryzen in performance though. It's fascinating that only
226.
by edg5000 6mo ago
:)))
227.
by edg5000 6mo ago
I think using the vision decoder baked into modern LLMs is the way to go. Have the LLM iterate; make sure it can assert placement qualities and understands the hard requirements. I think it can be done.
228.
by edg5000 7mo ago
Hahaha
229.
by edg5000 7mo ago
I build custom harnesses (like many of us) and I genuinely think Anthropic will eventually sue their customers if they detect they are selling competing harnesses (competes with their vertically integrated offerrings). I feel Alibaba and De
230.
by edg5000 7mo ago
What are the reasons for switching? Personally I got into the habit of doing a bit of a round robin with Codex/Claude (CLI) and then DeepSeek and Qwen web chat. And Claude in web chat. I like to switch just to learn the differences, ot
231.
by edg5000 7mo ago
Do they have an API where you can control the chat template or at least just put everything in the system prompt? This way you can control everything including the tool calling syntax. Even if you use the trained tool syntax, it allows you
232.
by edg5000 7mo ago
How can it be free? What do you mean? EDIT: Ah, I see. Some kind of promotion. Pretty cool.
233.
by edg5000 7mo ago
Has anybody done serious agentic work (e.g. using a CLI harness or simmilar) with 3.5 Plus/3.0 Max and such? How does it compare against Opus with Claude Code? I've used the chat quite a bit and I can't say at this point.
234.
by edg5000 7mo ago
They are repeating a million times on their huggingface page that the thinking output should be included in the conversation history for multiturn use. That makes me wonder, is this generally needed for LLMs? Because that implies that they
235.
by edg5000 7mo ago
The chat client has serious performance issues on lower end systems. Now I see why!
236.
by edg5000 7mo ago
We'd have to look at the longest-running democracies and observe how they handled periodic refactorings
237.
by edg5000 7mo ago
So limiting max context length also reduces VRAM needs a bit? If cache is 20% of total, 1/10th of context as a limit would mean 18% total memory reduction.
238.
by edg5000 7mo ago
For my desktop systems I do a nighly rsync to a central onsite server, which does nighly encrypted ZFS send to an offsite archive. I suppose a script could trigger on file change using Linux file watch features (exposed via CLI and C header
239.
by edg5000 7mo ago
What could happen is a reduction in the amount of programs used, with a smaller set of more sophisticated programs doing more work. This maps to what we saw pre-industrial revolution: lots of small family operations doing menial manufacturi
240.
by edg5000 7mo ago
This is very interesting. This could allow custom harnesses to be used economically with Opus. Depending on the usage limits, this may be cheaper than their API.

More ›