skerit

5 ms·

1.
by skerit 4d ago
At least you got them. I'm still at the "We will send you an email to do a final shipment confirmation once ready" stage.
2.
by skerit 4d ago
All they're saying to the other players is: let's lock it down and slice up the market between us. Claiming Skynet is just around the corner is a great way to get the public to believe only these huge companies can harness AI, and
3.
by skerit 9d ago
I never understood why any of the smartness had to be built into the TV in the first place. Sure, it's useful for the first year, and then the seriously underpowered hardware they installed into it will have trouble with just about any
4.
by skerit 12d ago
I use MCPs all the time, but just like in your case: they're my own MCP servers. They do exactly what I want and how I want it. They expose just the information and the features I want the LLM to have. And just like in your case, it co
5.
by skerit 20d ago
It is a bit weird. As someone from Belgium (we also use our own variety of Azerty) you are very frequently reminded that you're on a weird keyboard. At least for someone who games, you wouldn't make ZQSD the default. It's the
6.
by skerit 21d ago
So including me, you're like the fourth person in this section to get it on the second turn. That's peculiar.
7.
by skerit 21d ago
What the hell, same here. My initial guess was "CELLO", from there on I had 32 other possible words and somehow I got the right one?
8.
by skerit 23d ago
> The individual plans are heavily subsidized Not this again. There is not a single shred of evidence for this. In fact multiple times this year alone, people from Anthropic have said that inference and deployed models have positive marg
9.
by skerit 24d ago
> A week of using Codex more than Claude Which Claude is this about? Sonnet, Opus, Fable? All totally different beasts.
10.
by skerit 27d ago
I agree that "The person on the other side has the same tools you do" might be technically true, I have myself encountered it often enough that they don't even use it at all. People don't always want MY take on it, they
11.
by skerit 29d ago
Meanwhile every time their servers go down, your caches go down with it and you can pay for your entire context again. This sucks even harder when you like to use parallel subagents.
12.
by skerit 29d ago
It's been a while since the last reset. I think we're due one. Though I would prefer they just extend the +50% usage limit forever, it's been so long I can not imagine lossing a third of my current usage.
13.
by skerit 1mo ago
Does Claude's grep still prepend the relative path of the file before _every_ single line? Because that is nasty, especially in java projects.
14.
by skerit 1mo ago
Having Opus 5 on top of Fable is even more suspicious. Whatever this is measuring, I do not care. Opus 5 is shite compared to Fable.
15.
by skerit 1mo ago
I think one of the more interesting points is this: > Two-thirds of post-2022 data centers are located in water-stressed regions. I don't think it's wrong to ask questions about that _IF_ you also ask questions about all the ot
16.
by skerit 1mo ago
"That's on me."
17.
by skerit 2mo ago
Out of my 7 simultaneous sessions (my usage limit reset is tomorrow, so I have some lesser important projects to use my tokens on) there is still 1 session purring on. So there's at least 1 little Claude server still running.
18.
by skerit 2mo ago
And on her podcast with Vsauce she said the reason they chose that town was because their budget was so low and they could sleep over at a crew member's family there.
19.
by skerit 2mo ago
Interesting, they finally support `system` messages anywhere in a chat conversation: > Mid-conversation system messages are available on the Claude API, Claude in Amazon Bedrock, and Google Cloud. > > This feature is available on C
20.
by skerit 2mo ago
No matter how good compaction is, on some big projects it needs to read a lot of files. In my experience the first 200.000 tokens go FAST, but after that it slows down. Most of my Fable sessions don't go over 500.000 tokens, I don&#x2
21.
by skerit 2mo ago
I tried using `/goal` when it just came out, but back then they used Haiku for it. And if your main model is a 1M model, Haiku can't even read that much. So my /goal always failed. (I instead went for an elaborate /loop-
22.
by skerit 2mo ago
I think the improvement on how it codes is pretty much represented correctly by the benchmarks (a nice bump, but not some crazy leap) But where it really shines is in how NOT lazy it is. Fable requires less hand-holding. And I can understan
23.
by skerit 2mo ago
I've been doing something similar for some of my favourite older games. But the "byte for byte" claim has me worried. Isn't simply decompiling the sourcecode from the binary and releasing that problematic? It's not
24.
by skerit 2mo ago
I thought it was just Opus 4.7 and 4.8 that did this. Do other models do this too? Anyway: in my case Opus absolutely did not follow a similar instruction in the CLAUDE.md file. (But then again: it hardly followed _any_ CLAUDE.md instructio
25.
by skerit 3mo ago
That's not true, but it's their own fault for being so utterly terrible at communication. It only drops back to Opus for some prompts, just like before. The longer blogpost was a bit less vague about this.
26.
by skerit 3mo ago
Before Fable got pulled, they claimed their goal really was to keep it around in the subscriptions in some way, they just didn't know when it would return. This time there isn't even a single mention of this. It's just "
27.
by skerit 3mo ago
I just don't really understand the entire strategy behind this. Or their horrible, horrible communication. Because right now it's as if Fable/Mythos 5 is "the end of the line". It's as if this is the best their
28.
by skerit 3mo ago
This is why Fable was so good. It followed instructions and it was in no way lazy.
29.
by skerit 3mo ago
But LLMs aren't just shuffling words around. Words (tokens) are the "human interface" at the edges. Text input gets embedded, then there's this huge latent-space computation in the middle, and only at the tail end does t
30.
by skerit 3mo ago
I immediately saw the Ghidra one and was thinking: huh?

More ›