Julian
@mealreplacer
Followers
16,359
Following
1,096
Media
651
Statuses
4,238
Explore trending content on Musk Viewer
#KAZZAWARDS2024
• 540610 Tweets
SAROCHA REBECCA IN KAZZ
• 456964 Tweets
NORAWIT WITH PAUL SMITH
• 126997 Tweets
Varane
• 37744 Tweets
インプレゾンビ
• 33691 Tweets
魔法少女
• 25290 Tweets
B1NI TOPS SPOTIFY
• 24610 Tweets
FOURTH x KAZZ 2024🥳
• 23554 Tweets
#Destiny
• 18211 Tweets
#うたコン
• 16437 Tweets
名誉毀損
• 14118 Tweets
モグコレ
• 13166 Tweets
#仰天ニュース
• 12828 Tweets
風の行方
• 11763 Tweets
いなば食品
• 10216 Tweets
Pinned Tweet
Stop scrolling!
You’ve been visited by esteemed philosopher Robert Long.
Comment “good evening Robert” if you are currently having a subjective experience.
105
4
184
[Italian guy gets pulled over for speeding]
Italian guy: I so sorry officer you a caughta me red handed, what is the punishment for such a crime?
cop: it's a fine
Italian guy: bellissimo🤌, molte bene, you are a so kind
59
4K
35K
I want a gigantic red button I can press that instantaneously blocks every single tweet like this
Google Bard just got a massive upgrade.
ChatGPT is now falling behind.
Here are 10 things you can do with Bard that aren't possible with ChatGPT:
605
3K
15K
120
275
6K
“hey baby, did you eat today? I’ve missed you so much :) im wondering if you could take the usb stick i ordered you from amazon, upload this file onto it, and then plug it into your computer at work? you’re so smart and your muscles are so big im sure you could pull it off ❤️”
41
77
1K
Did you know it took over 500 BILLION gallons of water to train GPT-4?
That’s nearly *ten percent* of the entire Pacific Ocean!
We are going to run out of drinkable water at this rate.
197
41
905
Sometimes I think about this paper and I get a little sad knowing that it’s exceedingly unlikely that I will ever be able to produce satire that is as good or instructive as it is
21
114
902
After over 4 years of waiting, tens of thousands of dollars in costs, and navigating through a web of bureaucracy and disfunction, I’m happy to say that I finally got my U.S. green card.
🇺🇸🦅❤️
44
7
891
“have you ever trained a neural net?”
“nope”
“ok, but have you ever done any sort of machine learning?”
“I have no technical ability. I know nothing about AI. But I’m decisive about the memes I find funny”
9
51
896
I trained a neural net to predict whether an up-and-coming fantasy author will end up being wildly successful, based on a few writing samples. It sounds really fancy but all it’s doing is predicting the next Tolkien
23
70
829
does anyone know if there is an English language version of LessWrong
25
21
711
My girlfriend and I recently had a tough conversation about what we’d do if our respective LessWrong karmas diverged by more than 200 percent. We both agreed it would make sense for the person with higher karma to ghost the other
10
7
569
Under-appreciated fact: ChatGPT was only released 67 days ago 😱😱😱
27
21
556
There is still a surprising amount of alpha in reading and actually internalizing the material discussed in this very short and clear essay
17
32
516
guys what the fuck I think Clong can write funnier tweets than I can
41
23
485
have we thought about trying to align AI by having it interact with, like, the chillest person we know? like think of your coolest friend. they'd just be hanging out and kicking back. maybe it would like us after that
25
12
469
gf got upset with me because I’ve been “acting distant lately” and so I accused her of regulatory capture (I don’t know what this term means, I’m just accusing everyone of it these days)
10
24
450
'hyperrealistic video of entire industry of professional critics and skeptics lifting goalpost. They're moving the goalpost, under great strain, to another location.'
Introducing Sora, our text-to-video model.
Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions.
Prompt: “Beautiful, snowy…
10K
33K
141K
10
34
440
My 6 y/o son asked his teacher about OS AI at school today.
“Everyone should be able to access AI.” The teacher said. “Remember, you can use AI defensively.”
“But what if the offence-defence balance favours offence?” my son asked.
The teacher was stunned.
Proud parent moment.
35
21
414
In 2020, it cost $5 million to train GPT-3.
Today’s cost? $250,000.
That means by 2026, you will *get paid* $4.5 million to train GPT-3 sized models.
16
12
406
It’s fascinating that this particular line of reasoning has managed to become so memetically popular despite how utterly nonsensical and foolish it is
39
27
395
"Why would you think AI will end up taking control?"
"We will give it to them"
"But instrumental convergence has some serious conceptual holes in.... WAIT WHAT?"
"We will give it to them"
34
36
394
My non-EA friends told me there’s this thing called a “job” where they basically do a bunch of work trials every week for the same company, and then get paid a grant by that same company every 2 weeks.
Has anyone heard of something like this before?
12
4
325
Yeah, I’m a BAYESIAN:
B
A
Y
E
S
I sometimes change my opinion
A
N
9
28
341
BREAKING: OpenAI has filed a patent to give their language models “the sensation of having toes”, in the latest push towards multimodality
28
9
339
Exciting life update: I’ve recently joined the AI governance and policy team at Open Philanthropy!
I’m excited to have the opportunity to work on ensuring that humanity’s future is a flourishing one alongside a group of brilliant and inspiring colleagues.
47
7
337
Going viral on Twitter is such a wild experience. With a large enough N you come into contact with people that are more bitter than anyone you’d ever come into contact with in real life.
10/10 would recommend, it’s pretty fun
14
4
326
*Bernie Sanders voice* it is not fair that some AI models have billions of parameters
2
26
323
A man who is doomy about AI goes to doctor. “Doctor”, he says, “we aren’t on track to solve alignment”.
Doctor says, “treatment is simple. Great alignment theorist Eliezer Yudkowsky is in town. Go see him for instructions”.
Man bursts into tears. Says “you are trying to solv—”
6
12
329
your daily reminder that snoop dogg has better takes on AI than a large chunk of people on this website
16
34
310
Many people apparently hate EA so much that they’ll publicly and confidently claim that it’s bad to provide anti-malaria nets to people at risk of contracting malaria, without an ounce of scrutiny or suspicion.
Being this credulous towards a single shitty paper is embarrassing
10
9
303
The plan? Make everyone realize how cost-effective bednets are by writing a bad faith tweet designed to bait everyone
4
8
289
Guy who procrastinates so hard he goes “eh I’ll wait until AI is good enough to do it for me”
21
12
288
mechanistic interpretability mfs when they find the “paperclips” neuron
5
8
241
rationalist guys be like “why is she mad at me. this wasn’t in the sequences”
12
14
243
girl at bar asks for my social media. i show her my google scholar profile, hoping she’ll be impressed by my 81 citations.
“lol, you only put your papers on arxiv?” she says.
before i can respond, a chad peer review enthusiast comes and starts talking to her.
i go home alone.
9
1
239
[on Tinder date]
girl: yeah so I totally get the ick when a guy doesn’t understand Eliezer’s LessWrong posts
me: *stands up* look I’ve had a great time tonight but I just don’t think this is gonna work
7
6
219
Twitter folks who agree with this claim that AI does not “reason”: what does this even mean? I’m befuddled
137
3
237
*describes actions that lead to bad consequences* you see, here is why consequentialism is bad
9
18
230
If your go-to explanation for any and all social phenomena is “corporate greed” or whatever, you are going to have a very poor model of the various incentives faced by people and institutions, and a bad understanding of how and why the world works.
9
17
215
me *approaching cute girl*: (Lord grant me the self belief of the most confident man in the world 🙏)
...
me *comical french accent*: AI alignment is simply a matter of trial and error. We will just decide to build the safe systems, and not build the unsafe ones.
10
6
215
torturing animals is cruel, illegal, and utterly barbaric — that is unless you scale it up by the millions and sell their flesh for profit
2
31
215
Those of you who think the trajectory of frontier AI development is obvious, predictable, etc: make your predictions now, out loud, otherwise I will consider you just as confused and ignorant as the rest of us, despite any confident ex-post claims you might make later on
The AI thing to watch in the next months is Google Gemini. It is likely to be the first AI to beat GPT4. That means it can give us a hint of the future: How much better is it? Do larger models really hallucinate less? Do big general models keep beating small specialist ones? Etc.
51
107
792
34
16
209
Ohhhhh my god you're "accelerating" by founding a GPT-4 wrapper startup? should we tell everyone? Should we throw a party?should we invite sama
16
4
207
LLMs are incapable of creating anything new or novel — for instance, GPT-4 always uses the same letters I use when I write words
14
4
206
EAs have once again shown how cultish they are.
I’ve heard reports from anonymous sources that EAs are living together in the same houses, to “save money on rent” and “have people to hang out with”.
They’re calling this arrangement “having roommates” — more EA jargon!
Professed tolerance is often skin-deep.
Living multiple adults under the same roof is apparently unacceptably weird.
10
6
150
10
3
206
guy: what’s the deal with “AI risk”… it’s literally just math. you guys are scared of linear algebra and code
everyone: holy shit this guy is super smart. see how he just reduces everything to small little pieces? that’s awesome man
8
2
198
EAs would try negotiating and reasoning with the joker if the opportunity presented itself
Joker: “I’m going to set Gotham on fire and burn and kill everyone for fun, because I’m fucking crazy.”
EAs: “That strikes me as unintuitive. What’s your model? Let’s find the crux.”
13
9
189
I really loved the recent 80k episode on preventing the release of novel pathogens
3
11
187
fun fact, I once posted a question like this on Reddit: why philosophy when we have science? everyone called me dumb, but one person mentioned Peter Singer’s work as an example of important modern philosophy. A few Google searches later and I was on my way to being bed net pilled
why study philosophy at all? once the "queen of the sciences" it has been, not surpassed, but rendered irrelevant by more specialized fields of inquiry (linguistics, neuroscience, etc.). however, to not have read Plato, Aristotle, the pre-Socratics, Lucretius, is a disadvantage
137
36
301
10
6
180
I really, truly, love this website. It’s trite, but I will never log off.
9
3
180
my thoughts are that this kind of gimmicky gotcha stir-the-pot content collectively makes all of us dumber and less capable of having serious conversations about a transformative technology that touches/will soon touch every facet of life
"A Taxonomy of AI Panic Facilitators"
A visualization of leading AI Doomers (X-risk open letters, media interviews & OpEds).
Some AI experts enable them, while others oppose them.
The gender dynamics are fucked up.
It says a lot about the panic itself.
Your thoughts?
123
193
762
13
9
181
we need our brightest mechanistic interpretability researchers on this ASAP
5
6
176
"If you don't release your model weights when your startup only has a few GPUs, you have no heart, and if you don't withhold your model weights when you've spent 7+ figures on a training run, you have no brain." - Winston Churchill
4
11
172
My cat telling me I have to log off and go to sleep when I’m half way through writing a QT dunk on Yann LeCun
9
3
167
Guy who hates central planning but thinks the US government should award precisely 85,000 H1B visas per year
4
11
159
dwarkesh secrit interview strategy: do research and ask good questions to interesting guests
10
3
167
They’re callin it “Dolly-3”, T. You don’t need to do any of that prompt engineering bullshit anymore. The ChatGTP computer shit does it all for you. They made a photo of a sad avocado seeing a shrink — it’s fuckin’ incredible. Can you believe that?
2
5
162
So much nuance is lost when reading a podcast transcript. for instance, in the podcast, you can hear the pain in Will’s voice when Ezra laughs at him and calls him a soy boy.
5
4
160
@ns_whit
pretty sweet place! I usually go there for my legal advice, but I’ve heard the sushi is ace
1
0
158
me when some mechanistic interpretability paper I don’t understand gets published, but people seem excited
5
7
151
@VictorTaelin
You've earned my follow. I don't know if I'll keep it, it depends on what your more typical posts are like, but this is such incredibly respectable conduct that I want to try.
7
3
378
5
0
156
3
0
155