edg5000

19 ms·

181.
by edg5000 5mo ago
If everything is Google then there won´t be any more competition. They already have their hands in way to many things. Laptop margins are thin. They could squeeze smaller players out of the market with a decade of dump prices, seizing contr
182.
by edg5000 5mo ago
Commercial, paid: You must pay money, you get access to service or a binary, or something in between. No source code released. Shareware: You get to download and use the binary for free. Parts of it are missing, you must pay to receive a co
183.
by edg5000 5mo ago
Purely out of curiousity, I see you are using an em dash. Did you use voice transcription or something? It looks hand-typed though. I'm confused.
184.
by edg5000 5mo ago
In China it's become well known that promising new companies will get an offer from either Alibaba or Tencent. In the US, it's probably simmilar. Everything that's out in the open can get acquired or simply copied. Maybe that
185.
by edg5000 5mo ago
You're absolutely right!
186.
by edg5000 5mo ago
You bring up a good point. There is a general lack of awareness of how much power we're giving these monopolists. As a kid, in school I was thought to be weary of drugs, STDs, pimps and other threats. This should be added to the list.
187.
by edg5000 5mo ago
I think AWS is liked is because when AWS started, being able to get a new VPS up in minutes was still quite unusual. Many hosts would require about 24hr, I suspect, for getting a new VM up. At least those are some experiences I had. But now
188.
by edg5000 5mo ago
What's bad about AI: - Vendors get to know everything about you - Chips are becoming more politicized; I fear artificial scarcity as with housing will be put on chips, driving up prices. - It causes a lot of centralisation. No, I canno
189.
by edg5000 5mo ago
> I've been using GLM pretty much exclusively last 6-8 months. I have access to Anthropic and OpenAI models and others. I always keep returning to GLM, it isn't the best Very interesting. What sort of tasks do you use it for, a
190.
by edg5000 6mo ago
I feel the reasoning might be tuned for hard questions and not agentic work. I feel it overthinks, good for a very hard question, not for small incremental agentic steps. In theory, disabling thinking and using really well formed instructio
191.
by edg5000 6mo ago
Has anybody used V4 hard, for the most challenging tasks (agentically, locally)? It's so hard to compare without putting serious time in it. Like spending a year daily with the model.
192.
by edg5000 6mo ago
I think AI will create an open source dark age. Gradually, we'll see a lot less new good open source code. A gradual shift back to the proprietary world. Simmilar to the 1950-1990 period.
193.
by edg5000 6mo ago
My policy is that I don't allow agents to access all code. Some of it is shielded behind bind mounts. Maybe this is a pathetic, artisanal (or ego-driven), reaction of mine to the inevitable. I allow them to work on about 90% of the cod
194.
by edg5000 6mo ago
Really not looking forward to a regulated software industry. It will cause a lot of gatekeeping and bureaucracy. It's one of those things that may seem good, but in practice, it's pure waste in every way imaginable. Will just lead
195.
by edg5000 6mo ago
Have you compared it against other coding agents? What is your general workflow with DeepSeek; do you write a spec and then have it implement and test? Very interesting to hear. Becuase your harness is adapted to DeepSeek, you probably prom
196.
by edg5000 6mo ago
So the embedding model is a fixed-size view on a arbitrarily sized work history (tool calls, natural language messages)? The model is like a summarizer, but in latent space? And not aimed to summarize, but trained to hold whatever is needed
197.
by edg5000 6mo ago
My guess (I will test this eventually) is that you set a window size (which may be the model limit, or lower to reduce input token costs), the harness then refuses to show items that don't fit. If the model emits a command to read a fi
198.
by edg5000 6mo ago
Now I see why Anthropic isn't too happy with third party clients. The clients may not be so nice to their capacity as their own client, which has the interests aligned with minimum token consumption. A tricky dynamic.
199.
by edg5000 6mo ago
What I'm thinking is: When the agent wants to open more files or open more messages, eventually there will be no more context left. The agent is then essentially forced to hide some files and messages in order to be able to proceed. An
200.
by edg5000 6mo ago
So we agree on a message system having potential. But why the vectors? In any case, interesting stuff.
201.
by edg5000 6mo ago
Yes, it will destroy most of the caching potential. On the other hand, the average context window needed to achieve the same type of task may be much smaller. This might make up for it. And with a better harness, fewer rounds may be needed.
202.
by edg5000 6mo ago
There is nothing wrong with the HTTP layer, it's just a way to get a string into the model. The problem is the industry obsession on concatenating messages into a conversation stream. There is no reason to do it this way. Every time yo
203.
by edg5000 6mo ago
> (...) writing a genuinely good harness with lots of context engineering and solid tool integration is in fact not that easy. This. They are after the harness engineering experience of the Cursor people, I'd assume the they want to
204.
by edg5000 6mo ago
We're talking about completely different things. I'm talking about creating an LLM in the open, with individual contributors contributing to training sets as well as portions of the training work itself.
205.
by edg5000 6mo ago
Or the API is overpriced. The concept of charging per tokens does not map well to the actual costs an AI company has.
206.
by edg5000 6mo ago
But the API is incredibly expensive. I calculated that I would have spent 3000 EUR the last month, a lot more than the 100 I pay now.
207.
by edg5000 6mo ago
Very interesting, thanks for sharing EDIT: It's completely different though. This is more of a commodities market/auction/inference broker mechanism it seems.
208.
by edg5000 6mo ago
I asked Opus through claude code to set up the best local model fitting my hardware and that worked well for me. I could run Qwen 74B or something at .7 tok/s on my 64GB DDR5 on CPU. Pretty cool. Useful for overnight stuff. (this actua
209.
by edg5000 6mo ago
Is a community LLM possible? We'd have code to dynamically construct the pre-training dataset and use P2P mechanisms to share the acquired dataset. It would involve peer-crawling and other mechanisms to allow many people to contribute
210.
by edg5000 6mo ago
This is often a good strategy, but it's not easy at all. If your long term plan is to develop a product, focusing purely on delivery of services means no resources for product development. If you offer services purely with off the shel

More ›