kingstnap

6 ms·

91.
by kingstnap 4mo ago
According to many benchmarks this model is straight up frontier level and Zai seriously cooked. Some of these numbers are incredible. Excited to see if this turns out to be a Open Weight Opus 4.5 or better.
92.
by kingstnap 4mo ago
Cursor is massively overvalued. But so is SpaceX so it all evens out in the end. > each share of Cursor’s common stock and each share of Cursor’s preferred stock outstanding immediately prior to the Effective Time of the Merger will be a
93.
by kingstnap 4mo ago
What AI really seems to be posed to do is make labour a lot less valuable and capital a lot more valuable. Running a government requires a lot of cash flow. I think its sort of inevitable that it's going to need to start coming mostly
94.
by kingstnap 4mo ago
Highly reliable supply chains to bet the entire future on :)
95.
by kingstnap 4mo ago
Your pricing is ridiculously cheap. Cars are expensive. But aside from that you really can't offer unlimited usage. You would need to offer quota. Maybe something like 30 km / day that rolls over into a limit of 300 km.
96.
by kingstnap 4mo ago
Well he explicitly calls for regulations that ban open weight models, which of course hugely threaten Anthropic's API business model. If we got a open weight model actually around as good as Opus 4.6 that would be extremely bad for the
97.
by kingstnap 4mo ago
Its hard to read the first half of this as anything other than regulatory capture propaganda. It really all ties together as: > AI has become a major commercial technology >Frontier AI models, like airplanes, should be required to go
98.
by kingstnap 4mo ago
Given that MiMo is as cheap as Deepseek ( previous discussion: https://news.ycombinator.com/item?id=48282814 ) multiplying that by 3x for ultra speed is still shockingly cheap.
99.
by kingstnap 4mo ago
The fact that this article seems to honestly recommend people run 5 different type checkers on library test suits really reflects the tacked on feeling of Python typing.
100.
by kingstnap 4mo ago
Next to no one would be using less than the subscription price given how expensive Opus API is.
101.
by kingstnap 4mo ago
What's the high risk? That you end up with useless code? I didn't mean that the Dav1d people should yolo vibe code Dav2d. My point was this this is a very interesting possible experiment since there is no existing Dav2d contaminat
102.
by kingstnap 4mo ago
For those that don't know about this. Phi was announced with a paper called "Textbooks are all you need". What they did was use GPT 3.5 and created synthetic textbook chapters and exercises. They also did some more interestin
103.
by kingstnap 5mo ago
This is something I had thought about some time back where I was thinking about the feasibility of somehow using the upper and lower registers inside a multiplier as general purpose storage for fun / seeing if you could make them more
104.
by kingstnap 5mo ago
OpenAI's Dota 2 adventures were super hype back in the days.
105.
by kingstnap 5mo ago
This seems like an interesting case to test AI agents on. Like we had weird examples like C compilers and Bun. This is a much more interesting example because its highly nontrivial. AV1 exists, Dav1d exists. Lets see AI take the AV2 spec an
106.
by kingstnap 5mo ago
The reduction is in cached inputs. I've commented about this before but many labs, except Deepseek and Xaomi now, absolutely scam you for cached reads. You are basically paying out the nose for a few seconds of VRAM residence if you ar
107.
by kingstnap 5mo ago
I'm surprised they don't dwarf them by orders of magnitude. You can make arbitrarily random ETFs with all kinds of weird positions. Crypto, bonds, derivatives of all kinds, international stuff. Plus you can give them funny names l
108.
by kingstnap 5mo ago
You named the two biggest platforms [0], YouTube and WhatsApp are the social media. This is kinda like asking if Saudi Arabia and Russia are petrostates lol. [0] https://en.wikipedia.org/wiki/List_of_most_popular_soci
109.
by kingstnap 5mo ago
A significant caveat is that there is a pricing mismatch that makes it so first party's can subsidize quite heavily. Agents are expensive in large part because tool calls require round trips. It's because these APIs are stateless
110.
by kingstnap 5mo ago
Jevons paradox is at play. Right now frontier AI is very expensive which heavily suppresses demand. If you made it 10x cheaper right now you would see a truly unimaginable wave of bot slop.
111.
by kingstnap 5mo ago
Anthropic's caching requires you to pay a $0.75/Mtok for Sonnet and $1.25/MTok for Opus as a surcharge on top of the original input token cost. It's not even automatic. If you are reading ~8 times (8 total back and forth
112.
by kingstnap 5mo ago
It is animated but the viewer is broken for some reason (tested Chrome latest windows). This one works: https://www.svgviewer.dev/s/04ipQgsU
113.
by kingstnap 5mo ago
Exactly, clicking on the buttons is so slow. I want to be able to use keyboard keys.
114.
by kingstnap 5mo ago
I just looked at and you weren't kidding. There are significant changes every few minutes 24/7. https://github.com/openclaw/openclaw/commits/main/ GitHub insights over the last week. Excluding
115.
by kingstnap 5mo ago
A basic Google search leads me to this article [0]. > On March 27, 2026, Start9 CEO Matt Hill hosted a private unveiling of StartOS 0.4.0, the next major version of the operating system that powers the Start9 Server One. During that same
116.
by kingstnap 5mo ago
Terence Tao had a nice talk from the Future of Mathematics conference posted yesterday [0] that shapes a lot of my own feelings on this matter. The short of it is he argues how first to correctness shouldn't be the only goal / isn
117.
by kingstnap 5mo ago
I agree 100% with this article. You need to spend time on coding without agents and writing without AI as practice if nothing else. You should not get complacent in offloading all detail oriented work to agents.
118.
by kingstnap 5mo ago
The only thing that can actually introduce competition in RAM is some form of government backing around national security concerns. China has been doing this for some time though so there will probably be major Chinese supply coming in the
119.
by kingstnap 5mo ago
In Canada all news gets some amount of government grants and funding. Plus there are various tax programs and breaks related to news to help support them. People complain that it makes them biased but I don't really think so. At least
120.
by kingstnap 5mo ago
Yes Deepseek V4 pro is currently on discount. > The deepseek-v4-pro model is currently offered at a 75% discount, extended until 2026/05/31 15:59 UTC. However even when the discount ends its still very cheap. It will go back to

More ›