actually probably the most serious discussion of ai impact on employment i have seen
it is from bill kristol. as a millenial i am obligated to note how weird it is that i am saying anything nice about bill kristol every time i mention bill kristol
14.02.2026 02:10
π 223
π 25
π¬ 14
π 3
This Too Shall Pass Ψ§ΫΩ ΩΫΨ² Ψ¨Ϊ―Ψ°Ψ±Ψ― - Epic Iranian Music
YouTube video by Farya Faraji
www.youtube.com/watch?v=YFJ5...
10.02.2026 05:02
π 1
π 0
π¬ 0
π 0
[leftistly] "trump doesn't care about the law! he can do anything!! we've already lost!!!"
meanwhile
10.02.2026 01:50
π 580
π 119
π¬ 5
π 3
I can very much relate to this
08.02.2026 22:17
π 1
π 0
π¬ 1
π 0
I am not involved in research or advocacy in this area. Sorry.
24.01.2026 03:22
π 0
π 0
π¬ 0
π 1
Globally, seasonal influenza kills an average of 700,000 people each year from respiratory disease or cardiovascular disease.
During large flu pandemics, when influenza strains evolved substantially, the death toll was even higher.
21.01.2026 16:53
π 31
π 9
π¬ 1
π 2
What aspect of MIRI do you think is most useful? My impression was that they weren't doing much research anymore and have mostly switched to advocacy, and I'm not sure if they are good advocates. I might be wrong on all this, though.
23.01.2026 06:24
π 0
π 0
π¬ 1
π 0
I grew up in an area that was very diverse and included quite a number of Muslims. Rhetoric like this is always bizarre to me, almost surreal. Every American Muslim I've met has just been an American who is Muslim.
19.01.2026 21:53
π 1
π 0
π¬ 0
π 0
Erdos problems, a set of famous difficult math challenges, are a clear example of AI models breaching a threshold. The idea that an AI could solve one, let alone many, would have been insane a year ago (o1 was brand new). Now we have multiple Erdos problems solved by GPT-5.2 in the last couple weeks
18.01.2026 03:38
π 133
π 24
π¬ 5
π 3
Regulating supplements to the same purity/accuracy standards as drugs would weaken the woo market substantially while still allowing people to access actually-useful supplements (melatonin, etc.) It completely kills homeopathy if they have to prove the amount they are claiming is in the stuff
19.01.2026 07:15
π 1
π 0
π¬ 0
π 0
ICE can't even deal with irate middle-aged midwesterners. how does he occupy hundreds, if not thousands, of polling cites and precincts? trump v. illinois clarified that he has no legal authority to unilaterally commandeer national guards, how does he move forward from there?
15.01.2026 16:32
π 4493
π 455
π¬ 42
π 28
Also, we should offer statehood to the US Virgin Islands as well.
14.01.2026 05:58
π 3
π 0
π¬ 0
π 0
Why not three states if they want it?
14.01.2026 05:57
π 0
π 0
π¬ 0
π 0
A bus driver congratulated my mom on her honesty about my age. He said he'd have just claimed the under-13 discount if he were her. I was 16.
12.01.2026 21:11
π 1
π 0
π¬ 0
π 0
Important video. These guys are rattled. They can't go anywhere, they can't get anything done. They killed someone and it just got worse. They don't have any other intellectual or law-enforcement tools. They don't know what to do.
12.01.2026 01:59
π 922
π 174
π¬ 21
π 6
World βmay not have timeβ to prepare for AI safety risks, says leading researcher
AI safety expert David Dalrymple said rapid advances could outpace efforts to control powerful systems
Watching the AI safety debate unfold feels odd. Weβre talking about doubling times of 8 months for core capabilities. That's not a gradual shift; it's a phase transition. And the current public discourse feels built for gradual shifts... π§΅
www.theguardian.com/technology/2...
12.01.2026 15:43
π 2
π 1
π¬ 1
π 0
LLMs are rapidly improving. They are much better tools now than they were even a few months ago. It's easy to get out of date on their capabilities if you're not actively trying to keep up.
metr.org/blog/2025-03...
12.01.2026 20:04
π 2
π 0
π¬ 0
π 0
Those are unusually bad by llm standards because they are instant. There's a trade-off between speed and accuracy/usefulness. If you want more accurate answers, you want thinking models. You may want to try out the gemini thinking model and compare
12.01.2026 19:56
π 2
π 0
π¬ 0
π 0
I don't know about suicidal ideation or csam but there are various tools and metrics for evaluating hallucination frequency. Here is a hallucination leaderboard on hugging face using one such tool ranking frequency of hallucinations when summarizing a document
huggingface.co/spaces/vecta...
12.01.2026 19:42
π 2
π 0
π¬ 1
π 1
the best antidote to blackpilling is remembering that every video you see of an ICE thug abusing their power was recorded by an American who volunteered to put themselves in harm's way for the sake of people they've likely never met and never will
09.01.2026 17:40
π 4663
π 1130
π¬ 15
π 10
What's this a reference to?
03.01.2026 02:54
π 4
π 0
π¬ 1
π 0
Some good news in this:
* very effective anti-malaria drug!
* non-opioid painkiller!
* gonorrhea vaccine!
and more
17.12.2025 22:52
π 8
π 2
π¬ 0
π 0
What's wrong with balconies over the sidewalk? That's just providing more sun shade
07.12.2025 22:21
π 1
π 0
π¬ 1
π 0
afaict, a lot of leftists are fond of the letter x. I've seen "chicana" spelled as "xicana," and once even saw "xicanx"
27.11.2025 06:36
π 2
π 1
π¬ 2
π 0
MTG attacking Trump is great but actively bringing her into the Democratic tent risks accelerating the normalization of antisemitism.
18.11.2025 02:13
π 3
π 0
π¬ 0
π 0
what a pretty baby!
18.11.2025 00:57
π 2
π 0
π¬ 0
π 0
(This is something lots of rationlists themselves will say)
(background: I used to be a rationalist. I no longer consider myself one, but I still have many friends in the subculture)
17.11.2025 02:08
π 2
π 0
π¬ 0
π 0
I wouldn't say "standard California cult" because "cult" has the connotations of being a high-control group (which rationalism is extremely not), but I would say that it's fair to call rationalism a new religious movement grown in a Jewish substrate
17.11.2025 02:07
π 2
π 0
π¬ 1
π 0