nullbio

11 ms·

271.
by nullbio 2mo ago
Protobuf getting LSP support before Codex implements LSPs is wild.
272.
by nullbio 2mo ago
GPT 5.6 has always communicated autistically. I wish they'd put effort into fixing that.
273.
by nullbio 2mo ago
Until now the only uproar I've seen about it has been people posting Tiktok videos of them cutting down cameras or burning them with lasers, ergo commiting crimes.
274.
by nullbio 2mo ago
> If the model were just repeating the distribution baked into the training data, could you help me understand why the R^2 values in Table 3 are mostly below 0.1? Because it's only part of the picture. Firstly, the complexity metric
275.
by nullbio 2mo ago
It's very poorly executed and full of assumptions.
276.
by nullbio 2mo ago
Well, yeah, that's how LLM's work. They're next token predictors. If you give it sequences of text that are worded in a particular way, the sequences that follow will follow that statistical distribution from the training dat
277.
by nullbio 2mo ago
It is. That's why I said "If they have better accountability in terms of law enforcement using it (warrant requirements, transparency, etc) I see no problem with it." To be clear, I am not in favor of Flock in its current inc
278.
by nullbio 2mo ago
The US in the 70's and 80's is not comparable to current US given different socioeconomic factors. China is far safer than the US is, if you want an accurate example. > Is that why prisons so safe and free of crime? Surveilled
279.
by nullbio 2mo ago
99% of the people who attend protests take their phone with them. Furthermore, Flock does not prove you participated in a protest. What if you were going to a shop near a protest? What if you were going to the park? This would never be cons
280.
by nullbio 2mo ago
Great quote, but I don't consider this "essential liberty".
281.
by nullbio 2mo ago
And what percentage of that backlash is by people with a criminal record? And what percentage of these criminals are the ones pushing the narrative of this being some grave invasion of privacy that must be stopped for "mah freedom&quot
282.
by nullbio 2mo ago
OpenAI is run by its employees and their culture is far healthier. Unlike Anthropic, they don't have a CEO that actively encourages their employees to be dishonest.
283.
by nullbio 2mo ago
I agree with that too. So if the argument is: "Flock is good in principle but right now it doesn't have the necessary safeguards", well, that's exactly what I said in my original post.
284.
by nullbio 2mo ago
Now imagine that none of that is relevant because we're not talking about listening devices or communication tapping. This is just false equivalence, and this false equivalence is why people are immediately so defensive of anyone who s
285.
by nullbio 2mo ago
Nice hyperbole. I'm concerned about govt listening to our private conversations and mining our inner dialogue. I could not care less if they know when my car passes a camera in public. People are instantly very emotional about this wit
286.
by nullbio 2mo ago
Flock is not tracking you 24/7, your phone is. What is so dangerous about the government knowing that your car passed a camera at a particular point in time?
287.
by nullbio 2mo ago
Right, and I think that part of Flock should be fixed. But I don't think in general it is a bad idea. Do these same people also think it is bad that the police can get a warrant to read your text messages?
288.
by nullbio 2mo ago
What are you so concerned about exactly?
289.
by nullbio 2mo ago
https://student-cms.prd.timeshighereducation.com/sites/defau... https://student-cms.prd.timeshighereducation.com/sites/defau... https://student-cms.prd.timeshighereducation.com/sit
290.
by nullbio 2mo ago
It would indicate that there is a circulatory issue in general that causes this. If so, there may be other blockages as well, at various other lymph nodes that connect to the brain, which would explain why it works better for some and not o
291.
by nullbio 2mo ago
Does it matter if it's just temporary? If it works, it is the smoking gun of what alzheimers is and why it happens. If it is not a permanent fix, the next question is what is happening in the body that is causing these blockages. It al
292.
by nullbio 2mo ago
If this works it will be buried so that the pharmaceutical industries can find a pill intervention, no doubt.
293.
by nullbio 2mo ago
Ask it to do a zebra on a jetski to escape the benchmax.
294.
by nullbio 2mo ago
Yeah, but here's a dirty little secret that very few people are discussing: You can't use Claude for this sort of thing if the goal is to make better AI systems. Anthropic finetunes Claude to dissuade people and the agent from usi
295.
by nullbio 2mo ago
This isn't a great argument. Researchers who do not think string theory is good are not going to spend years on it. There are plenty of experts that dismiss string theory. I have no skin in the game, and don't care either way, fwi
296.
by nullbio 2mo ago
Because the model you get access to and the model that they use at Anthropic and sell to businesses with million dollar contracts are entirely different, despite being named the same. Anthropic is one of the most corrupt companies on Earth
297.
by nullbio 2mo ago
They don't need to slow down development, they just need to slow down on being greedy over all of the hardware. They should be investing in manufacturing.
298.
by nullbio 2mo ago
We need more manufacturing capacity, or this will never end.
299.
by nullbio 2mo ago
No one will ever convince me that the Mac approach for installing programs is intuitive or good UX. The whole argument of: "But if you think about it in ... way, it makes sense" is a terrible argument.
300.
by nullbio 2mo ago
Have you done any benchmarks on this approach? I wonder if it just ends up in reward-hacking, or if it actually substantially changes the quality of the code. Are these static analysis metrics enough on their own to cause a substantial impr

More ›