perching_aix

5 ms·

61.
by perching_aix 1mo ago
It's hideous. It's puzzling to me that the two things were made by the same person (LLM caveat nonwithstanding).
62.
by perching_aix 1mo ago
> Don't tune for speed until you've measured, and even then don't unless one part of the code overwhelms the rest. Genuine question, is software performance really linear like that, that one can and should only fight the t
63.
by perching_aix 1mo ago
Not necessarily a "risk". Some people treat their agents better than they do their colleagues. Sometimes the colleagues have it coming, sometimes not. Sometimes it's one's own fault. Agentic interactions can help you
64.
by perching_aix 1mo ago
But there are already voice models that do a reasonable job at a fraction of the throughput available? The real question is how expensive it is to coordinate between these different modalities, and I really don't see why it'd be a
65.
by perching_aix 1mo ago
No, they don't. Or at least didn't. See what the news article is about...
66.
by perching_aix 1mo ago
> At low concurrency scenarios, Jalapeño demonstrates remarkable interactivity, hitting over 700 tokens per sec per user at concurrency 1 on the DeepSeek R1 model. This is about the same rate you get out of Sol Ultraspeed. Why do you thi
67.
by perching_aix 2mo ago
I thought we could have a good thing go on. If the judgement is already cast, why let it go to waste?
68.
by perching_aix 2mo ago
Yet nevertheless, both e.g. syntax highlighting and intellisense have been heralded as things that "fry ur brain". So have high level programming languages by the way. To give you a particularly specific example, Rob Pike really w
69.
by perching_aix 2mo ago
So does this one.
70.
by perching_aix 2mo ago
I do really enjoy the style of their blogposts, they remind me of the Cloudflare postmortems. Wish their models could produce it. Not sure I agree about e.g. GPT not sounding this formulaic though, imo it definitely does. I don't actua
71.
by perching_aix 2mo ago
And what are we supposed to do with the two of you apparantly living under a rock [0], and going through baby's first "kids these days" moments, exactly? Never heard of LMGTFY ("Let Me Google That For You")? Why do
72.
by perching_aix 2mo ago
> Is this the zoomer syndrome I am hearing about lately? Sounds easily searchable...
73.
by perching_aix 2mo ago
Can you give an example? It's a bit entertaining to read this given all the years long (and still ongoing) posturing about LLM sycophancy (not that all of these couldn't be true at the same time).
74.
by perching_aix 2mo ago
Do architects lay bricks often? Wasn't aware.
75.
by perching_aix 2mo ago
I have been as well. Based on my own sessions, Max vs Max, same 1M context window size, the literal majority of the cost overhead of Opus vs Sonnet comes from Opus being chattier. So I started using Low reasoning instead of falling back to
76.
by perching_aix 2mo ago
> network portable They're an absolute torture to work with over a dodgy connection, which often may not even be your fault. Looking at you AWS.
77.
by perching_aix 2mo ago
If this wasn't just some cheap insult, it's worrying that you think one can become licensed for arbitrary confidence like that. https://en.wikipedia.org/wiki/Nobel_disease
78.
by perching_aix 2mo ago
Around the clock?
79.
by perching_aix 2mo ago
I notice that a lot of these terms are squarely humanist for certain people, so any kind of allegation that machines are exhibiting them as traits will be a complete showstopper for them. You'll either be strung along on an infinite go
80.
by perching_aix 2mo ago
becomes rampant? Where have you been the past decades? Hell, centuries and millenia? If we're going cynical, may as well go full throttle.
81.
by perching_aix 2mo ago
Given the 750 tok/sec GPT 5.6 Sol Ultrafast (via Cerebras), the many-1000 tok/sec Chinese models, and the 15000 tok/sec Taalas HC1, I think we're well on the way towards seeing that solved too. Combine the two, and yeah,
82.
by perching_aix 2mo ago
Sounds like a variation on the induced demand principle: https://en.wikipedia.org/wiki/Induced_demand Related: - Parkinson's law: "Work expands to fill the available time." https://en.wikipedi
83.
by perching_aix 2mo ago
Oh neat, cheers. Credit to GPT 5.6 Sol for actually unearthing this on my behalf, too. :)
84.
by perching_aix 2mo ago
Let me know if it worked out in the end, would be cool if it did.
85.
by perching_aix 2mo ago
4) it's fun and quaint, like an odd, jazzy version of a player piano from an alternate universe
86.
by perching_aix 2mo ago
The site owner me doesn't work with incessantly helpless people. Good for them. I do. "They all have the same tools as I do" is exactly what's infuriating. They either blatantly don't use them, or are too incompeten
87.
by perching_aix 2mo ago
Didn't know macOS is this "literally 1984", it's a bit wild. It got me to apply some slop of my own. Supposedly there's some new provisioning scheme that'd allow the author to not have to disable SIP: https:&#
88.
by perching_aix 2mo ago
How do you mean?
89.
by perching_aix 2mo ago
In the sense that you could block the model from doing specifically that, yes. The issue is, fighting the model like that doesn't scale. It has to figure out on its own what's expected, that's where the whole utility of it al
90.
by perching_aix 2mo ago
It's strange, because I have zero doubt in my mind that Claude could 100% genuinely reimplement the driver and port it to actually natively run on macOS, were it prompted that way.

More ›