pyaamb

3 ms·

1.
by pyaamb 2mo ago
'model welfare' as a concept seems so premature that I can only question the motivations behind pushing for it at this particular point in time
2.
by pyaamb 2mo ago
On the flip side, its important to know just how much we would be sacrificing if big frontier gets their way in convincing the courts that distillation is a bad thing
3.
by pyaamb 2mo ago
I really need the spacebar to expand slightly more on the right so my thumb doesn't miss it while typing. Wish it were customizable.
4.
by pyaamb 2mo ago
+1
5.
by pyaamb 2mo ago
I would like that. I wonder if its something a software update can enable or not
6.
by pyaamb 2mo ago
Need this but for AI models. Sleep mode where they self prune and optimize overnight for their daily purpose.
7.
by pyaamb 3mo ago
The Pelican Brief
8.
by pyaamb 3mo ago
Also perhaps AI agents are now capable enough to run these apps the way the user would and recognize these dark patterns. Flag those and feed it back to a warning at the point of sale that users can upvote there to signal their disapproval
9.
by pyaamb 3mo ago
I guess it falls on how the courts see it. I'm sure the app devs would love to take their money.
10.
by pyaamb 3mo ago
Does explicit licensing and the threat of a lawsuit related to violation of it not help here?
11.
by pyaamb 3mo ago
Need to rethink the system that allows for (and encourages) this kind of plausible deniability. From "Oh we need this permission for [non essential feature] and you need to accept it if you want the app at all" -> to giving the
12.
by pyaamb 3mo ago
Wow.
13.
by pyaamb 3mo ago
I prefer the Browser OS test
14.
by pyaamb 3mo ago
Between humans, I feel that what we like to call a "good communicator" as opposed to someone who just rattles off facts or prepared statements comes down to the "theory of mind" skill and how advanced that is. The presen
15.
by pyaamb 3mo ago
In thinking of directions where LLM's could develop from here, I cant help but think that a models ability to self introspect would immensely improve their utility. The R&D on how to achieve that is beyond me though. How do you tra
16.
by pyaamb 3mo ago
A necessary tool to manufacture consent in the feudalist society we're speed running towards
17.
by pyaamb 3mo ago
Elon may have been technically correct about 'if humans can do it with vision only then theres no reason why vision-only cannot work' but turned out to be the wrong strategy for learning how to get there. lidar gives you a constan
18.
by pyaamb 4mo ago
Do you see any negativity towards llama.cpp? This isnt drama-seeking. The author makes very valid points.
19.
by pyaamb 4mo ago
Its about closing the gap. its the gap over everyone else that will give one country leverage over everyone else in the AI age. Makes me wonder what the world would look like if a country or group of countries did this during the industrial
20.
by pyaamb 4mo ago
That is seriously impressive. Can I ask the prompt you gave it?
21.
by pyaamb 4mo ago
Data centers should at the very least build their own renewable energy generation. It would set the right incentives in place and encourage investment in clean energy generation solutions. It would also present a very compelling problem to
22.
by pyaamb 4mo ago
the ignorance is a consequence of our societies system of incentives. everything keeps pointing back towards that each time theres an event like this. But if we aren't going to attach an economic price for emissions then it remains an
23.
by pyaamb 6mo ago
editor themes seem like a good candidate for something that someones trusted local LLM could generate for them
24.
by pyaamb 6mo ago
we are caught between an evolutionary need to know that our existence is meaningful and a universe that seems indifferent.
25.
by pyaamb 6mo ago
If its offline I love it. Otherwise I hate it.
26.
by pyaamb 6mo ago
a VPN enabled wifi router would suffice as a fallback tho right?
27.
by pyaamb 6mo ago
when can we hold lawmakers personally responsible for any consequences resulting from passing bad laws?
28.
by pyaamb 8mo ago
power right where its needed, plus shade for your vehicle. this is the way.
29.
by pyaamb 9mo ago
I would argue that: 1. "slop" doesn't come from just AI. Take a look at the headlines on your daily newsfeed. Carefully crafted by humans. 2. This is an application of AI that serves human values by what it attempts to preser
30.
by pyaamb 9mo ago
I've been casually daydreaming about possible paths an "AI augmented" newsfeed experience could take. I'm imagining an initial group of agents scraping the web for events (twitter/x, science/eng/business p

More ›