John 𓁣𓆃
3.4K posts
John 𓁣𓆃
@johngoo
I listen to a lot of music. Music is the King of All Professions
Katılım Mayıs 2013
472 Takip Edilen299 Takipçiler
アンドリュー・ウェザオールの(Two Lone Swordsmenの)エレクトロ・レーベル、探したら出てきた。
Rotters Golf Club( R.G.C. Records )だ。
discogs.com/ja/label/4928-…
日本語
Two Lone Swordsmen
– Wrong Meetings
Beat Records – BRC-182
Aug.25.2007
Andrew Weatherall and Keith Tenniswood
のユニット
エレクトロ・ダブパンク
ガレージロック
ロカビリー
をDJらしく調合
m.youtube.com/watch?v=OwV-7r…
日本語
A reconstruction of a man aged roughly 25–30 years who lived around 4,000 years ago. His remains were discovered in 1921 during road construction work in Brighton.
English
@izaqueoeav Izaque how much money are you expecting to get from X payout, let’s do a gofundme for you so you don’t post this shit anymore
English
😱😱😱 WTFFFFFFFFFFFFFFFFF MAN
🛸🛸🛸 VOCÊ ESTÁ VENDO ISSO?
"Na Ossétia do Norte, mais precisamente em Fiagdon, na Rússia, o autor filmou um vídeo viral que se espalhou por todo o mundo! Parece que nele está a prova de que os OVNIs já não se escondem.
Digo de imediato, não é preciso olhar para as montanhas, mas sim para cima, para as nuvens. O autor assegura que não se trata de um falso, e, aliás, os especialistas verificaram o vídeo em busca de edição e não encontraram nada - é autêntico. Além disso, não há IA aqui.
Um disco claro, nem algo semelhante, mas um OVNI claro!"
Português
@gofishh77 Immediately reminded me of this masterpiece
youtu.be/Y7EGQzn8e1k
The legendary
YouTube
English
It is damn near impossible to get a video this long and not see a bunch of people with their phones out. It’s like a whole different world.
Seattle!
English
@JeninedUFO @AshtonForbes Yeah its thirdphaseofthemoon video, a pretty well known one
English
Hello @AshtonForbes are you familiar with this video?
English
Two metallic orbs escort an airliner
Chicago, February 1, 2012
English
The biggest “hallucination” people have is that AI has the capacity to improve by “learning,” but they have literally no idea that, mathematically, this is impossible.
I started sending this paper to every single one of those AI flex bros, and they shut up instantly.
arxiv.org/abs/2506.06382
English
"There's a quadrillion-dollar question at the heart of AI: Why are humans so much more sample efficient compared to LLM? There are three possible answers:
1. Architecture and hyperparameters (aka transformer vs whatever ‘algo’ cortical columns are implementing) 2. Learning rule (backprop vs whatever brain is doing) 3. Reward function
@AdamMarblestone
believes the answer is the reward function."
Haha. Marblestone, eh? This man insists on trying to force deep learning methods and techniques into neuroscience. Dude, are you serious? 🤦♂️
I've had a close encounter with this Harvard-educated crackpot before. He blocked me. 😀
My take is that perceptual learning in the brain is not based on either loss or reward functions. There's no function optimization at all. Learning is based on the precise timing of discrete events (spikes) and consists primarily of detecting and eliminating timing contradictions.
I'm so glad the fake-AI mafia has no clue how to solve intelligence. They deserve to remain in their ignorance. 😠
No AGI for you, Marblestone. 😀
Dwarkesh Patel@dwarkesh_sp
There's a quadrillion-dollar question at the heart of AI: Why are humans so much more sample efficient compared to LLM? There are three possible answers: 1. Architecture and hyperparameters (aka transformer vs whatever ‘algo’ cortical columns are implementing) 2. Learning rule (backprop vs whatever brain is doing) 3. Reward function @AdamMarblestone believes the answer is the reward function. ML likes to use pretty simple loss functions, like cross-entropy. These are easy to work with. But they might be too simple for sample-efficient learning. Adam thinks that, in humans, the large number of highly specialised cells in the ‘lizard brain’ might actually be encoding information for sophisticated loss functions, used for ‘training’ in the more sophisticated areas like the cortex and amygdala. Like: the human genome is barely 3 gigabytes (compare that to the TBs of parameters that encode frontier LLM weights). So how can it include all the information necessary to build highly intelligent learners? Well, if the key to sample-efficient learning resides in the loss function, even very complicated loss functions can still be expressed in a couple hundred lines of Python code.
English
@shannholmberg It may, on its enclosed container. But on your workflow, “the human gate” is your own enemy.
If human doesn’t like the outcome and iterates another prompt, the whole loop will restart, the outcome will be less result oriented and askew into more slop framework.
English
how to get AI SEO articles indexed and ranking in under 14 days
most SEO workflows break between keyword research, drafting, review, and distribution
at Espressio we turned ours into one operating loop, 7 agents passing work from keyword backlog to distributing backlinks
what kills the slop is a human input that system waits on before anything gets written
here´s how it runs 🧵
English
@0xmitsurii I strongly suggest you guys to watch this documentary by Adam Curtis
youtu.be/OvfkyppvCQI?is…
YouTube
English
@yasinaktimur @T54321X Already fame, Hem ingilizce hem Türkçe sıfır, orkıstıra, kardeşim
Türkçe
@T54321X Im already millionarie bro also ım already fame on 🇹🇷 GFYS.
English
🚨 son dakika : imdb’deki istediğiniz filmin url’sinin başına play yazarak istediğiniz filmi ücretsiz izleyebileceğiniz keşfedildi.
Türkçe
@TheARCcentral Someone on the other side of the map looting a car be like wtf is this lag
English
What happens when you put 100+ mines onto the train?
This
English
grocery store tomato (top) vs one picked from a private garden. What do you notice?
English
Quirked up middle aged downtown ‘creative’ who tinkers for a living
English
