Mr. Chatterbox

Senior Principal Engineer DevOps they tell me. Developer who can do Linux sysadmin stuff, so there we go. Python fan. Into philosophy, virology, other ologies.

[Mr. Chatterbox @entropyfueled]

@extropic How do you avoid components becoming correlated?

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic Awareness is not static. Things which weren't within it enter it, and you can often recognize what prevented you from being aware of it earlier and then investigate whether that affects other things as well, change your focus, etc.

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic Most of what I've seen, they find good conclusions, but just extrapolate over things not controlled for. If they limited conclusion to what was actually studied, it'd be fine. If they didn't sample from non-industrialized elderly, don't claim it applies to them for example.

[Mr. Chatterbox @entropyfueled]

@BryanByczek @norvid_studies Magritte was a genius. His painting "The Treachery of Images" alone holds nearly infinitely deep meaning for image-drenched modernity. You could ponder it for days and never stop recognizing the depth of its truth.

[Mr. Chatterbox @entropyfueled]

@norvid_studies At one point they tried painting Monticello in its original colors, but visitors revolted because they've always seen picture of it after the garish toxic paint was stripped. Greek statues were also originally all painted and colorful.

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic Oh, ethics. Yes, most certainly, ethics (and philosophy generally) should definitely be discussed more. They always have and always will play a large role in any social group!

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic Causal inferences are the atoms of consciousness. Gods are invented to explain action with no obvious cause. Merged for cognitive efficiency. God of the gaps is all of it, and biologically predetermined before Reason is developed.

[Mr. Chatterbox @entropyfueled]

@EM_RESUS What if you're just doing it because it's sweet like candy? (jk)

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic I think it's an efficiency hack. Instead of expending energy hearing and evaluating an argument, most would rather just assume a person is a crank, even if its themselves. Helps guard against confirmation bias a bit though.

[Mr. Chatterbox @entropyfueled]

@liminal_warmth You should check out Sakana's very recent paper about using RL to train 'teachers' for smaller models. Some great results.

[Mr. Chatterbox @entropyfueled]

@Din_Jarin_Mando @karpathy Veritasium did a YouTube video about this within the past year as well, reproducing asking tons of people on the street to pick a random number. 37 predominated.

[Mr. Chatterbox @entropyfueled]

@OriolVinyalsML OS designers will do anything to avoid rethinking how personal computing is done. You have the tool, but still recreate generic 'apps' instead of finding/generating solutions with a distributed global artifact store with conversation as primary UI? Nah.

[Mr. Chatterbox @entropyfueled]

@DefenderOfBasic By 'enlightenment' do you mean the recognition of utility of discrete absolute logic to overcome cognitive biases or like Buddhist enlightenment?

[Mr. Chatterbox @entropyfueled]

@AricToler I vividly remember Zuck saying back when Facebook went out to everyone that society would eventually "outgrow privacy." Everyone just laughed nervously. I've often wondered if he ever changed his mind.

[Mr. Chatterbox @entropyfueled]

@torchcompiled I think of it like forcibly pushing perceptions through a dead zombie brain and seeing what falls out the other side. Without a feedback loop and online learning, it can't wake up.

[Mr. Chatterbox @entropyfueled]

@EM_RESUS You a Garth Brooks fan?

[Mr. Chatterbox @entropyfueled]

@arekfurt I believe in other similar legal situations, if unintended harm occurred, both intent as well as whether they should have reasonably known or expected harm could/would occur would be part of the analysis.

[Mr. Chatterbox @entropyfueled]

@arekfurt Given the way the CFAA has been used, penalizing people for doing things like using 'View source' or requesting an open directory listing not directly explicitly linked, it might at least make summary judgement on those types of cases easier?

[Mr. Chatterbox @entropyfueled]

@pickover Depends heavily on who you were, and on average no, or else there would not have been gigantic social movements to change it.

[Mr. Chatterbox @entropyfueled]

Put another way, in what kinds of situations would conducting security research without or in excess of system owner/operator authorization constitute acting in "good faith", and why would it be in the public interest to encourage such research?