samrus

5 ms·

61.
by samrus 3mo ago
I need good handwriting OCR. Like perfect. Because id love to write things like blog posts on a remarkable or something, and then be able to just export it to text and publish
62.
by samrus 3mo ago
It sounds like extortion. Give us your face. It'd be a shame if something happens to all that data
63.
by samrus 3mo ago
Model allignment isnt a science, its a craft. And its very very imprecise. Getting anything that subtle through would be impossible without making it obvious And if your still concerned, go ahead and have a non chinese model review the code
64.
by samrus 3mo ago
You have the wieghts for the chinese models. Go ahead and run whatever tests you want. The american models on the other hand are too powerful for simpletons like us to be allowed to use, but we can be assured its all in our best interest, c
65.
by samrus 3mo ago
That is so true. These models need to be democratized. The chinese are doing that because its beneficial for them in second place. The americans are trying to stop it because its detrimental for them in first place. Simple incentive allignm
66.
by samrus 3mo ago
> OpenAI's internal "Galaxy" model is fully capable of functioning as what security people refer to as an "Advanced Persistent Threat This is marketing. They saw anthropic create crazy hype around (the admittedly grea
67.
by samrus 3mo ago
Cowards. Force them to implement resealable waterproofing
68.
by samrus 3mo ago
What i absolutely love is that this exact strategy is why coding agents dont work with proprietary tools. Open standards have so much context out in the open that any harness/agent can work effortlessly with it. But things like excel a
69.
by samrus 3mo ago
> how is development "not transparent and open" in an open source project? Im guessing cathedral rather than bazaar?
70.
by samrus 3mo ago
Legal issue no, but with how much more likely cops are to be domestic abusers, i think its very dangerous
71.
by samrus 3mo ago
It basically shows a lack of respect for the gravity of these tools. If a cop was fucking around with a gun, youd want them held accountable for that so there is a standard of care in using these things. We need that because these can be ve
72.
by samrus 3mo ago
Sysiphus be like
73.
by samrus 3mo ago
> The value of philosophy outside a self-referential intellectual game is extremely dubious I wouldnt go that far. I think your clutching at straws a little bit. Its a real stretch from philosohers are insecure to they are useless. This
74.
by samrus 3mo ago
People do complain about bureacracy with the police. They dont phrase it that way. What they say is that when they need the police, its never able to help. People dont realize that in addition to the police being corrupt and lazy, its also
75.
by samrus 3mo ago
Its a tradeoff between efficacy and accountability. Its tough. What access syrem would you propose? Genuinely asking
76.
by samrus 3mo ago
Reminds me Kai Lentit's ffmpeg video https://youtu.be/9kaIXkImCAM
77.
by samrus 3mo ago
In my opinion, most iltra rivh get to where they are by being narcissictic sociopaths. But zuckerberg really takes the cake. Remember when he thought a makeover and singing listens would be enough to make him president? Mans actually derang
78.
by samrus 3mo ago
Thats not as complex or scalable as what LLMs can do. The capability and scale is what changes the equation. Quantity is a quality unto itself here
79.
by samrus 3mo ago
Its game theory so perfect play is assumed. Spherical cows and all that
80.
by samrus 3mo ago
Not really. They arent actually using more resources this way either. This might be a fundamental inefficiency thats being removed It kinda makes sense too. Because while people do read code word by word, we often "glance over" it
81.
by samrus 3mo ago
I swear if the ghouls running things had abit more decency and allowed people to actually access and controll their passkeys then that would be the future, everyone would adopt it. The experience is so nice with key pair exchange for ssh. I
82.
by samrus 3mo ago
They redistributed the statistical patterns of those copyrighted materials. Which perhaps should be treated similarly nos As for your "technically not copyright infringement" defense. Those laws are from a time when those patterns
83.
by samrus 3mo ago
> You cannot do that in the hopes of viewers donating $5 here and there. But then how does twitch work? Genuinely asking, whats the difference?
84.
by samrus 3mo ago
It is
85.
by samrus 3mo ago
What your describing is iterating in the last mile. Your assuming that the AI will iterate in the last mile with the same efficacy that it iterates before that. I think that will fail. Thats the bitter lesson, that adhoc solutions to the la
86.
by samrus 3mo ago
Yeah. I think LLMs wont be able to do that on their own or with hueristics. We'll need to bake game theory into the base model for that
87.
by samrus 3mo ago
Bro is the training data
88.
by samrus 3mo ago
> but in fact are building piles of tech debt that will fall over before long This is speculation as well. Its well founded but speculation nonetheless. Youre speculating things will stay the way they have till now. I do see your point b
89.
by samrus 3mo ago
Tracks. This is that same speculation that AI will be good enough. Wonder what a crash in that isecase will look like? Increased suicide rates? More instances of psychosis? It might not wven be directly measurable or easily traceable to AI
90.
by samrus 3mo ago
Yeah but the investments arent aiming for churning out SaaS apps. Its to automate large swathes of intellectual labour. Of which only SWE has been cracked yet. There is a question mark as to if the others will crack. If they arent then thes

More ›