Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
walrus01
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
29 ms
·
631.
▲
by
walrus01
3mo ago
Sure seems like it. Real humans don't write exclusively one sentence comments. It doesn't look like the post history of anyone here that I'm fairly sure is a real meat based mammal. It looks like someone gave an LLM an instru
632.
▲
by
walrus01
3mo ago
All aboard the public relations hype train! choo choo!
633.
▲
by
walrus01
3mo ago
> Why not just use claude Cowork? Because people want to be able to do things like use their own clients of pi or opencode with LLMs they run themselves, such as the just released deepseek v4 flash 0731, not permanently tied to an Anthro
634.
▲
by
walrus01
3mo ago
This really isn't a tailscale specific compromise at all. If your private key store or credential store to any tool is compromised, particularly if those credentials can be used to automate standing up new network connections/se
635.
▲
by
walrus01
3mo ago
Fucking over Bunnie Hwang would be a stupid move on an epic scale considering how well known he is.
636.
▲
by
walrus01
3mo ago
the unsloth GGUF at <165GB will run on most 256GB RAM pure-CPU systems (or with llama-server and a mix of loading as much as you can onto a single 32GB, 48GB or 96GB GPU and the rest onto system DRAM). https://huggingface.co&#
637.
▲
by
walrus01
3mo ago
From 12 years ago https://theonion.com/fbi-uncovers-al-qaeda-plot-to-just-sit-...
638.
▲
by
walrus01
3mo ago
another link: https://futurism.com/artificial-intelligence/data-center-noi... This particular datacenter was retrofitted into an existing industrial building, there was some sort of industrial metalworking factory ther
639.
▲
by
walrus01
3mo ago
Brought to you by Carl's Jr.
640.
▲
by
walrus01
3mo ago
I'm not so sure that allowing AI agents to interact in a way that's actually indistinguishable from a human sitting at a keyboard/mouse is a great idea. What I wrote above will likely become technologiclly possible, but it&
641.
▲
by
walrus01
3mo ago
> Due to the limitations with browser and computer use capabilities, Saul could not post on platforms like Reddit and Product Hunt. At some point in the future with a LOT more tokens and speed, it'll be possible to give a tool a ful
642.
▲
by
walrus01
3mo ago
My personal theory on that is that you bought all your stuff at a specific time (like 2005 or prior) before the population of original 1980s NES-game-owners became old enough, established in their careers and started having enough disposabl
643.
▲
by
walrus01
3mo ago
Googling "psa graded 9.6 NES mario bros" turns up some absurd looking dollar values. I can totally see collecting old video game stuff as a hobby when we're talking about things that a teenager can get for free or nearly free
644.
▲
by
walrus01
3mo ago
Something that I wonder about when I see the 'final' results like these NES cartridges sealed in a graded plastic box as a collectible. Surely the market value of these will only hold up as long as they're within the normal l
645.
▲
by
walrus01
3mo ago
You can also try a reply like "please send back a photo of yourself with a shoe balanced on top of your head" and see how they reply. The replies from a human who is going "wtf?" will be very different from an LLM.
646.
▲
by
walrus01
3mo ago
Unsloth studio is just llama.cpp/llama-server under the hood, so you should see the same performance with the latest daily compiled llama-server (and the right CLI options to load a GGUF file) and any of your own choice of tooling on t
647.
▲
by
walrus01
3mo ago
Sometimes I wonder how much of Russ Hannemann from HBO's "Silcon Valley" was based on real life anecdotes of people who interacted with Cuban. RADIO ON THE INTERNET!!!
648.
▲
by
walrus01
3mo ago
> Edit: and there it is, the post has been flagged, to be kept off the front page and suppressed from discovery. it wasn't flagged because of some nefarious plan to suppress people from knowing about this, it was flagged because it&
649.
▲
by
walrus01
3mo ago
You'd have to ask Anthropic or OpenAI, but I've seen probably more than a dozen reports of it happening. You also risk getting your account banned for discussing certain risky things in chemistry, biochemistry, weapons, conflict,
650.
▲
by
walrus01
3mo ago
> I don't care who King Charles is every single time There is just a stupendous amount of everything packed into a 35B size or larger model. For instance Qwen 3.6 35B A3B (Q8) can do a fairly decent job translating English to Arabic
651.
▲
by
walrus01
3mo ago
Quoting from the article: > In our runs, K3 served 16 concurrent sessions (GLM-5.2 managed 24). Aggregate token throughput is about 30% lower (122 vs 170 tok/s at 16 users), and median task time is about 50% longer (38 vs 26 minutes
652.
▲
by
walrus01
3mo ago
So the 'unsloth' Kimi K3 Q8 quantization is about 1.56TB which is close to the original safetensors file in size, and should be nearly lossless. Realistically with enough context you'd need 2TB total RAM whether it's run
653.
▲
by
walrus01
3mo ago
Keep in mind that lm studio is just a GUI with llama.cpp/llama-server under the hood, so if you want to compile the latest llama-server and load your own choice of model, then connect to it with pi, opencode, etc or your own other choi
654.
▲
by
walrus01
3mo ago
Forget five nines, our new goal is to be better than nine fives.
655.
▲
by
walrus01
3mo ago
The marketing copy on this website looks very suspicious, from the point of view of somebody who actually knows solar systems.
656.
▲
by
walrus01
3mo ago
500MHz was not a thing in 1998, an expensive desktop PC was a slot based Pentium 2 350MHz in 1998. The 500MHz P3 came at least a year later and was quite high end for its time. In early 1998 a more normal baseline home gaming desktop PC w
657.
▲
by
walrus01
3mo ago
I got a job at the local Uzbek restaurant, but before I could learn all the recipes and secret ingredients, I had to sign a naan disclosure agreement.
658.
▲
by
walrus01
3mo ago
At least it's not minutes/token?
659.
▲
by
walrus01
3mo ago
tokens/second, no, more like seconds/token
660.
▲
by
walrus01
3mo ago
The G3 iMac that came out in this time period had much weaker 3D rendering capability than a contemporary PC gaming video card like a TNT2. Their video ability was closer to the 2D/office oriented video cards sold for non gaming use...
More ›