Sabitlenmiş Tweet
bling
4.6K posts
bling
@blingdivinity
Artist, Open to Interpretation 𖣘 American Undergraduate 𓉱 Casting Mathemagical Spells 𝜆
East North Central Katılım Mart 2020
2.3K Takip Edilen1.1K Takipçiler
@minordissent always-blue pressers realize the problem with always-red is that "not everyone will just press red"
but they dont realize always blue has the same problem
"not half of everyone will just press blue"
its half as severe, but can still fail
English
@minordissent and the always-blues are the most annoying about it because they are focused on what others should press, which in their mind motivates them to evangelize for blue, which runs counter to the whole point of the scenario, since the scenario relies on no prior coordination!
English
It’s very interesting how all the blue pressers focus on what OTHERS SHOULD press, the low IQ red pressers focus on what THEY THEMSELVES SHOULD press, and the high IQ red pressers focus on what OTHERS WILL press.
Crémieux@cremieuxrecueil
Accurate cartoon. In my polling, almost 6% of people said they picked the wrong button when they were prompted about it. Assume the global error rate is higher and kids' responses are random, and easily 1-in-6 to 1-in-5 hit blue by mistake. So, if red wins, society is ruined.
English
@lumpenspace its only thinking if it comes from the precision-weighted hierarchical bayesian inference via reciprocal message passing minimizing expected free energy region of the brain duh
English
what are they doing when they relate various fact through analogy and transfer items of knowledge across domains in order to find the solution to a problem?
Christopher David@Tazerface16
People understand that LLMs aren't actually "thinking," right?
English
@lexrus yep cool work! very similar, slightly different approaches but overall converged on a few of the same tricks (e.g,. using skylight private apis for window focusing)
English
i reverse engineered codex background computer use
ai can click, scroll, type, drag/resize, and send key cmds - all in the background, while you keep using your computer
runs as a localhost api. scriptable. written in swift
video is realtime. github below
English
@andthatto can you please share some transcripts? i would love to see some output examples
English
Qwen 3.6 is frontier for local.
It also thinks forever.
I tried a dumb inference-time trick: make its block obey a tiny grammar.
Result:
- HumanEval+: 22x fewer think tokens, no accuracy loss
- LiveCodeBench public slice: +14% pass@1, ~5x fewer total tokens
English
@max_spero_ seems like the ad was written by ai. pangram for AR glasses when
English
@blingdivinity @XiaomiMiMo please don't submit an issue, this will stay as is.
English
People of pi.dev. As a weekend gift, we added @XiaomiMiMo Token Plan as a first class provider. I also made some breaking changes for the better. If you have custom providers and models, point pi at the changelog so it can fix them up for you.
This will be a recuring theme in the coming days and weeks. We'll get through it together.
English
@badlogicgames @XiaomiMiMo it would be great if the ui reflected provider specific reasoning effort names too. just went through my config and mapped claude effort but now i have to remember that pi xhigh is really opus 4.7 max, pi high is xhigh, medium is high, etc. i will submit an issue
English
the OpenAI goblin fiasco was a Big L for the interpretability research community
They solved the mystery without SAEs or probing or anything. just talked to various models and counted the number of times they said Goblin
English
@blingdivinity Right! Perhaps there will be musical agents. I.e., an entity creates the music like how entities do writing
English
yeah its a stupid thought experiment and fails for the same reason all thought experiments do: map and territory become irreconcilable. the structure of the imaginary world you created must have somehow caused the event of the thought experiment. but for purposes of the experiment you must try to ignore what those things could be. for your example: we could have a day where zero people in the world die in a world very different from ours. but that difference would be so large it would make trying to use the thought experiment to think about our own world impossible
English
@blingdivinity @IsaacKing314 It's not about IQ. It's just a statistical impossibility. It's like saying, could we have a day where zero people in the world die?
English
This has got to be one of the most failed collective action problems of all time
MrBeast@MrBeast
If this tweet has exactly 1 like in 24 hours I’ll give that person $1,000,000
English