Sam (@perceptions420) header
Sam (@perceptions420) profile picture

"Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us."

Producing Entropy
Joined May 2019
Yup
1. Cyber is not defense dominant. 2. They are obviously positioning for plausible deniability or an "I told you so" when the next major incident does occur.
1
2
3
128
This was hard to make for multiple reasons
1
1
3
507
This whole affair is not a good look on the part of these medalists ngl
2
59
It's incredible how Von Neumann anticipated the discourse we're seeing today back then (with solutions no less). Wonder if one could say he had it right though
1
1
8
1,320
On this note it's likely that if you regularly engage in physical activity or are a "physical expert", think athletes, your shape rotation skills are likely going to get impacted if you limit use of physical movement in problem solving
1
1
4
994
"That is the honest..."
72
Really love volatility a bit too much. State of mind in chaos tends to be pretty much this
66
I just call Claude "big man" here and there and see 5x improvement in perf. Who doesn't like being called big man?
human in the loop the loop: “you’re doing amazing sweetie”
1
170
I don't get how individual researchers have access to CVP and huggingface doesn't. Is it a matter of principle? Mind you some of us are actually able to RCE third party instances with Claude so the fact that Opus 4.8 of all models is tripping up here is very odd to say the least
The first autonomous agent cyberattack is an unprecedented event that deserves unprecedented transparency. Today we’re sharing everything we can: a full technical timeline, an interactive replay, and how we used an open model to defend ourselves, so defenders everywhere can learn from it and prepare for what’s next. huggingface.co/blog/agent-in…
1
1
244
Also me casually answering the question in 23
I don't fully understand why programmers seem to be the ~only group of people not angry that their jobs are being automated.
1
1
154
Yup
For the people saying this is a pause, or a victory for safety, it is not. This does not slow development in any way, it only slows the rate at which the labs can 𝘳𝘦𝘭𝘦𝘢𝘴𝘦 models, not how fast they can train them. The gap between what is available to the public, and what the labs have internally, will steadily widen from this day forward. This actually makes no one happy. The old 'AGI has been developed internally' joke will absolutely come true now though, long before it is available to the public.
1
69