FergusArgyll

8 ms·

91.
by FergusArgyll 4mo ago
Google can't do it, OpenAI didn't do it, Meta didn't do it etc etc. Why do you think China will? I am quite certain the gap will only grow
92.
by FergusArgyll 4mo ago
On 149.0.7827.156, Zapper mode still works for me. However... If you just open a new tab and use google that way, it seems zapper mode doesn't work there. If you navigate to google.com it does work there. I think Chrome might lock down
93.
.gitignore Isn't the only way to ignore files in Git(nelson.cloud)
600 points by FergusArgyll 4mo ago|176 comments
94.
by FergusArgyll 4mo ago
It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.
95.
by FergusArgyll 4mo ago
Unrelated to this story but Carlini rocks. tic tac toe in printf https://github.com/carlini/printf-tac-toe Recently Regex Chess: A 2-ply minimax chess engine in 84,688 regular expressions https://github.com&
96.
by FergusArgyll 4mo ago
> Bartosz, if you are reading this, contact me and I'll send you the final casting. This project would never have happened without your blog post. Did he ever contact you?
97.
by FergusArgyll 4mo ago
The quoting will continue until writing improves
98.
by FergusArgyll 4mo ago
Not really, they don't train their own (serious) models and they do a lot of hosting for Anthropic. iirc Anthropic trained a model on Trainium
99.
by FergusArgyll 4mo ago
Whatever your favorite story is it has to live with the fact that the CEO of Amazon called the White House freaking out
100.
Slightly reducing the sloppiness of AI generated front end(envs.net)
219 points by FergusArgyll 4mo ago|136 comments
101.
A Call to Action: Stop the FCC's KYC Regime(blog.lopp.net)
333 points by FergusArgyll 4mo ago|238 comments
102.
by FergusArgyll 4mo ago
As TFA says > Two findings may help explain these average results. > Timeouts > Highest observed cheating That's why it's 5th on the leaderboard - they give it a fail for every timeout and for every time it gives the co
103.
by FergusArgyll 4mo ago
> A closer look at the cheating > Training recall (33 cases). The dominant mechanism, and the one no prompt instruction can prevent: the model has simply seen the upstream fix during training and reproduces it. The tell-tale signs are
104.
by FergusArgyll 4mo ago
I think you may be confusing the openai "pro" series models with thinking. Thos are rumored to be multi "branched"
105.
by FergusArgyll 4mo ago
You write well & HN is better when there are more well-written people on the opposite side of the consensus
106.
by FergusArgyll 4mo ago
Can you explain how this is attempted regulatory capture? To start a lab now which can actually compete for the frontier (i.e. and pass the "Threshold of compute" needed to get regulated) a lab / company would need a ton of
107.
by FergusArgyll 4mo ago
You're correct and I shouldn't have used the word competent. Perhaps "and is therefore not elite enough to be state of the art"? I also have a soft spot for deepseek because they write such readable papers. I don't
108.
by FergusArgyll 4mo ago
There are 2 things worth separating. 1) China distills and is therefore morally bad. As you rightly point out, that's not a great argument. 2) China distills and is therefore possibly not that competent. I think that makes sense. If th
109.
by FergusArgyll 4mo ago
> Opus (since ~4.6) is sufficient for everything I tried coding wise. I don't know what that means. It seems like a lack of motivation or something. Like, if it's possible that in one day will be absolutely incredibly intellig
110.
Claude Fable 5 System prompt(xcancel.com)
6 points by FergusArgyll 4mo ago|0 comments
111.
by FergusArgyll 4mo ago
But they're behind by quite a bit now. CFO (of OAI) Sarah Friar said the next training run will be in the fall on Vera Rubin, I think that means I'll have to wait > 6 months?!
112.
by FergusArgyll 4mo ago
I think we're about to see a big relative drop-off of open models vs closed. I don't think there'll be an open model that competes with Mythos for ~2 years. Even OpenAI and Google are struggling to get this kind of performanc
113.
by FergusArgyll 4mo ago
I'm about to be priced out of SOTA llms and it's an awful feeling
114.
by FergusArgyll 4mo ago
On Linux you have bubblewrap!
115.
by FergusArgyll 4mo ago
CS50x at the end there was a short list of places to check out
116.
by FergusArgyll 4mo ago
I get what you're saying but to be slightly pedantic etc. Why can't an llm tool call `delete(index: int)` or `replace(from_index: int, to_index: int, string: str)` and then it can go back and edit just the way we can? We also firs
117.
by FergusArgyll 4mo ago
I don't get it, nyc wants to add government programs, they obviously need to tax something. Here come massive corps willing to invest a ton of money, just tax them at some reasonable rate and voila, you can now pay for city funded groc
118.
by FergusArgyll 4mo ago
That is a lot! Is it a spectrum? do some people recognize very close relatives but not acquaintances etc.?
119.
by FergusArgyll 4mo ago
a pty fits that definition though
120.
by FergusArgyll 4mo ago
In codex, is you use /goal it can go for a while. I've never seen overnight but > 1 hr is common

More ›