solenoid0937

6 ms·

91.
by solenoid0937 2mo ago
Dario doesn't have a problem with open weights, he just thinks that open weights should be tested prior to release so you aren't giving everyone a zero-day button or a "make a virus" button. That's it. That's t
92.
by solenoid0937 2mo ago
I use Kimi K3, Opus 5, and Fable on a daily basis. Fable is the only one that reliably one shots complex changes and makes the right design choices. Everything else requires handholding. I can let Fable loose on a 12+ hour (for AI) task and
93.
by solenoid0937 2mo ago
Not much. 0.04% of US water production. For context, that's about 3% of the water consumption of our golf courses, and less than 1% of what we use to water our lawns.
94.
by solenoid0937 2mo ago
White collar work will require SOTA models up until the point where said models can automate white collar work entirely. Then, we might finally see a meaningful commoditization crunch.
95.
by solenoid0937 2mo ago
This is just pure histrionics, you have no understanding whatsoever of the scale of data center water consumption, electricity consumption, new capacity coming online, etc.
96.
by solenoid0937 2mo ago
> If you are just vibe coding, you need to use the most capable model you can find and even then it might not be good enough I mean, this proves my point. Better models enable you to get more done. With Fable, 80% of the time, I no longe
97.
by solenoid0937 2mo ago
This stuff is just obvious to anyone working with the latest models. Fully autonomous agents are a game changer. Having to pair program with one is indeed "last gen", I haven't done that for a month and I won't ever be d
98.
by solenoid0937 2mo ago
Unless you are very cash strapped, Fable is a very nominal fixed cost compared to the benefit of what it offers (fully autonomous agents, and no longer needing to pair program with one). And even it isn't "enough". I can very
99.
by solenoid0937 2mo ago
I work for a FAANG, and have my own personal projects for which I use the Chinese models. and the big models do indeed save/make us a lot of money. The Chinese models are not good enough for anything other than pair programming, whic
100.
by solenoid0937 2mo ago
I think HN doesn't really understand fixed costs, I spend a few hundred dollars per day on Fable and the costs are irrelevant compared to what we make.
101.
by solenoid0937 3mo ago
They are communicating that they have reached AGI in order to create a trail of documentation for the Microsoft legal battle
102.
by solenoid0937 3mo ago
I use K3 a lot, it is far behind Fable imo. For personal projects I use Kimi and for anything that matters I use Fable.
103.
by solenoid0937 3mo ago
Opus 5 is amazing if you use it as an autonomous agent as opposed to a pair programmer. It's like Fable at a much lower price.
104.
by solenoid0937 3mo ago
It seems pretty obvious that the solution is auto mode (running a classifier on each action) + sandboxing
105.
by solenoid0937 3mo ago
Surprised no one is talking about auto mode, it solves this problem.
106.
by solenoid0937 3mo ago
Neat. I didn't know psychologists could cure terminal cancer!
107.
by solenoid0937 3mo ago
Capital hoards benefits because resource scarcity and cost of production are a thing Post scarcity societies would likely have entirely different dynamics
108.
by solenoid0937 3mo ago
I think death, and most extreme pain/suffering should be optional. Superintelligence, if done safely, lets us solve most of our problems.
109.
by solenoid0937 3mo ago
HN was definitely using "stochastic parrot" throughout most of 2024 and a good half of 2025 to dismiss AI capabilities.
110.
by solenoid0937 3mo ago
> I’m partly in that camp. LLMs are intelligent and general, which are the two conditions of AGI. Others believe that we’re building a god in a box, and that it’ll doom all of humanity. It’s hard to take a field seriously when the basic
111.
by solenoid0937 3mo ago
I don't trust them, they are probably distilling on your data.
112.
by solenoid0937 3mo ago
No, it's not just people calling themselves smart, it is a specific philosophy of how to think. Whether you think that philosophy works or not is another matter. IMO it has its flaws but is far superior to vibes-based hot takes you see
113.
by solenoid0937 3mo ago
certainly, it has enabled the buildout of cities and infrastructure that were previously impossible to build.
114.
by solenoid0937 3mo ago
It's a rationalist community, IMHO their methodology of thought leads to much more well-reasoned takes on AI than on HN, where the discussion here is often very emotionally charged or led by wishful thinking. When it comes to AI, LessW
115.
by solenoid0937 3mo ago
Skeptical
116.
by solenoid0937 3mo ago
Most of HN does not take the idea of superintelligence seriously, and until this year did not take the idea of AGI seriously. I think LessWrong is a much better community for rational takes on AI, they've been reasoning about these r
117.
by solenoid0937 3mo ago
Everyone knows there isn't a moat. No one joins OpenAI/Anthropic unless they think these companies will reach superintelligence. So most people joining believe their equity will 10-100x even from where it is today. (Coincidentally
118.
by solenoid0937 3mo ago
Google does not have to compete at the frontier, they already own a lot of Anthropic. It's not an "issue" for them because it's not one of their goals.
119.
by solenoid0937 3mo ago
I have a friend at Google DeepMind who tells me that Google believes in AGI/superintelligence just as much as HN does, which is probably why no one with conviction wants to work there.
120.
by solenoid0937 3mo ago
Google owns 14% of Anthropic. If Anthropic gets to superintelligence, Google wins as well. Google doesn't really need a frontier lab of its own.

More ›