cootsnuck

6 ms·

31.
by cootsnuck 4mo ago
> something like the universal approximation theory comes to mind, transformer architecture clearly has the shape of a universal algorithm approximator That says nothing about emulating a human brain.
32.
by cootsnuck 4mo ago
I'm hearing different from PhDs. The bottleneck with much research isn't "trying out ideas" so much as it's all the bureaucratic minutiae, grants, mentoring PhD candidates, collaboration with other researchers, etc.
33.
by cootsnuck 5mo ago
I call it "I'm not like other girls" writing.
34.
by cootsnuck 5mo ago
> 4.8 is also 2x more expensive for a "modest" performance bump. How refreshing. Where are you seeing it's 2x more expensive? https://platform.claude.com/docs/en/about-claude/pricing
35.
by cootsnuck 5mo ago
I'd be kind of shocked if a model that came out six months ago is the same size and cost to run as one that just came out today.
36.
by cootsnuck 5mo ago
Well, it seems like collectively we are all struggling to perceive model progress, given that it seems like every reply to you is reporting different experiences with which of the models has subjectively performed best for them.
37.
by cootsnuck 5mo ago
I have my original gmail address from 20 years ago still and even old youtube videos my friends and I uploaded from ~18 years ago. The cringe is rough but at some point the cringe becomes so bad it loops back around to me just feeling nosta
38.
by cootsnuck 5mo ago
Also it was but a few months ago that their CFO said, in a court filing, that Anthropic's revenue across the entire lifetime of the company "exceeds $5 billion". Pretty strange. https://www.reuters.com/comment
39.
by cootsnuck 5mo ago
Yea, I wonder the same. Something I've been wanting to read up on is how the land ownership works with these data center deals. Because developers promising massive projects to scoop up a bunch of land, do little to nothing with it, si
40.
by cootsnuck 5mo ago
Yea, my bad. "majority getting scrapped” was sloppy wording. More accurate to say would be that a big chunk of the AI data center pipeline looks delayed/speculative. 16GW is slated for 2026, but only 5GW is actually under construc
41.
by cootsnuck 5mo ago
I'd argue we don't need to build a 9 GW data center anywhere. But that's just me.
42.
by cootsnuck 5mo ago
> The data centers the industry wants are all going to get built. If that was the case then why are the majority of data center projects getting scrapped? https://gizmodo.com/data-center-project-cancellations-quadru... W
43.
by cootsnuck 5mo ago
Heat. 9 GW puts off a staggering amount of heat. In an already arid environment. What's the worst that can happen? https://www.sltrib.com/news/environment/2026/05/07/utahs-dat...
44.
by cootsnuck 5mo ago
It seems fairly likely that a comically gargantuan data center like Stratos would endanger the surrounding ecosystem, at a minimum. https://www.sltrib.com/news/environment/2026/05/07/utahs-dat... Bu
45.
by cootsnuck 5mo ago
No, they do not have empathy or motivations. Arguably, if you think of them as having such then maybe it could help you coax out better outputs occasionally (wildly dependent on the task at hand). But that's only because of the LLM alw
46.
by cootsnuck 5mo ago
But the firms in the headlines doing layoffs after layoffs aren't growing crops, selling groceries, or driving busses... They're knowledge work roles in companies selling intangible products and services. It's large corporati
47.
by cootsnuck 5mo ago
Yea. LLM inference requires batch processing to have a shred of hope at being cost efficient. Batch processing requires a not so insignificant amount of scale (but probably not as much as people think). I'm very pro local models, but n
48.
by cootsnuck 5mo ago
Not my term! Some real academics came up with it: https://www.normaltech.ai/p/new-paper-towards-a-science-of-a...
49.
by cootsnuck 5mo ago
I work specifically in voice AI and am very familiar with how these tools and systems work. > I would expect an "AI Note Taker" to faithfully transcribe the entire conversation. With the same quality I see in a lot of automate
50.
by cootsnuck 5mo ago
Yup, spot on. There's a capability-reliability gap that the industry does not like to talk about too much. It often feels like the AI industry is continually glossing over the fact that capability and reliability are fundamentally diff
51.
by cootsnuck 5mo ago
They absolutely need deterministic tools. What you just described is exactly how the current popular AI agents work. They use "harnesses", which to me is just a rebranding of what we have known all along about building useful and
52.
by cootsnuck 5mo ago
Yes! I was just about to comment the same thing. I sank so many hours into that Dragon Ball Z game. Was called Dragon Ball Z Tournament. And its background music was an instrumental version of Sisqo's Thong Song. Wild. https:/&#x
53.
by cootsnuck 5mo ago
I think "AI-tone" is a much better way to characterize this stuff than accusing people of using AI. The problem has always been the same. Putting out slop feels disrespectful to the people you want to read/watch your stuff. M
54.
by cootsnuck 5mo ago
Their comment does not give off "LLM-written" really... It drives forwards actual points without superfluous segments. I don't think it's helpful to try and discredit people whenever we want by throwing around accusation
55.
by cootsnuck 6mo ago
Sam, we're not going to use your weird eye scanning orb.
56.
by cootsnuck 6mo ago
Because you can't have a tech company offering third party identity verification solutions if you just go with something like an RTA header.
57.
by cootsnuck 6mo ago
About the friction, not the capabilities...I haven't switched off my biz calendar/appointment provider I'm paying for even though I've kinda outgrown it. I wouldn't under estimate switching friction.
58.
by cootsnuck 6mo ago
Yup, you're absolutely right. The open source models do have their rough edges. I use NVIDIA's Parakeet v3 model a lot locally, and it will occasionally do this thing where it just repeats a word like a dozen times.
59.
by cootsnuck 6mo ago
Handy has Windows support. https://handy.computer/
60.
by cootsnuck 6mo ago
Yup, Handy is the one that made me stop looking for local open source alternatives to Wispr Flow. I'll give a shoutout as well to Glimpse: https://github.com/LegendarySpy/Glimpse

More ›