Comment by wg0
11 hours ago
If you're an investor you should try Deepseekv4 before you put your hard earned money in this gambling spree.
Context - Deepseekv4 is freely available to download you can host your own and sell it keeping the proceeds and it rivals Claude Opus 4.7.
"Thank you for your attention to this matter"
It is worse than GPT5.5 ... but is so cheap that it doesn't really matter.
cheap for whom? local hardware + electricity bills don’t even begin to get close to frontier model subscription even in price
> Deepseekv4 is freely available to download you can host your own and sell it keeping the proceeds and it rivals Claude Opus 4.7.
good luck getting a machine that can run its specs though. Even flash is goign to require ponying up 5-10 grand to run the minimal specs for it. The vast majority of people will find their machine falls behind as tech progresses long before they get a return on that investment. That said, it does mean there will be a healthy market for "generic providers" in the AI landscape with these open weight models.
It's not "there will be", there already is a market of generic providers and you can use millions of tokens of DeepSeek-v4-flash for like 1 dollar.
https://models.dev/?search=deepseek-v4-flash
Yeah, but will those hosted models help me write smut, advance my weekend CBRN hobby, advice on how to kill myself, advice on how to kill the person who made me want to kill myself, and how to set up a mega drug manufacturing operation like a real life Walter White?
3 replies →
You only need about a mac w 96GB or 128gb to run deepseek v4flash with ds4(https://github.com/antirez/ds4). Works mostly well
That's only antirez's 2-bit version though. The real version of DeepSeek V4 Flash will be slow on that machine.
Companies still pay 5k and more for a basic website. 5-10k is quite affordable.
Investment is not basic math. Its also dependencies to US companies, trust etc.
Investment and finance in general, at this scale, is far more geopolitics than it is math. It’s self evident.
We’ve all seen how the “math” on so much of the AI business sector literally doesn’t check out, and here there are: still ballooning, still making deals, still directly crafting laws through political influence, still taking over damn near every user space.
Politics at a high enough level lets you play a different game with different rules.
Politics at a low enough level lets you do the same actually, but we usually call that civil unrest, guerrilla warfare, or collective action depending on how many of which group is defying which “rules”.
It’s very easy to get used to the guardrails and guidelines around us when they persist and succeed for decades, but they are much more fragile than they appear.
https://github.com/antirez/ds4
> good luck getting a machine that can run its specs though.
That's any machine that can physically host the weights and context. You'd need a highly-specced machine for better performance and throughput, but it's not a requirement as far as literally executing the model and getting output.
Can I rent space in Colossus2?
But...so does the tech sector. They will also have to continually upgrade their AI slop data centers to run newer better models, generating a heap of waste along it. And that money has to be made back.
Even investor knows this well, they will still do that if they can make other people to invest their hard earned moneys into this after them.
Isn’t knowing how to scale and optimize llm traffic the main barrier ?
That's just a "more hardware" problem.
Right, and who has more hardware?
3 replies →
I'll bet Deepseekv4 could answer any questions you had related to that. How much of a moat will it prove to be in the long run? "Scale and optimize" sounds like a commodity business.