llm_trw

37 ms·

661.
by llm_trw 2y ago
>1. Ukrainian President Volodymyr Zelenskyy says North Korea has become directly involved in the war, with a senior Ukrainian intelligence official saying around 3,000 North Korean troops are now in Russia, and are training for deploymen
662.
by llm_trw 2y ago
Do you pay attention to the people who win at Russian roulette more than those who lose? OxyContin was marketed to be the safe alternative to all opioids that came before it, impossible to become addicted to and extremely difficult to overd
663.
by llm_trw 2y ago
Every decade or two I hear this about some wonder drug only for them to turn out just as bad as what they were replacing. The opioid epidemic being the latest one.
664.
by llm_trw 2y ago
Welcome to the 80s part 2. Hope you enjoy the ride.
665.
by llm_trw 2y ago
>We also paid to make the Nature Aging paper open access. What a world to live in.
666.
by llm_trw 2y ago
Providing toilets to employees does not tie to profit. Toilets in all offices are now closed. We are saving over $10k per day.
667.
by llm_trw 2y ago
Doesn't matter what the granularity of your proof step is you still need to understand what that proof step does to be a mathematician. You seem to mistake syntactic brevity for semantic meaning.
668.
by llm_trw 2y ago
No. It's pretty common to have depth 20+ on even relatively simple proofs when you start looking at decomposing proofs to their axioms. The old joke about 1 + 1 = 2 being on page 360 of Principia Mathematica still largely holds.
669.
by llm_trw 2y ago
Option 3 anyone who can add Ai to their career with be even more highly paid because of the productivity boost.
670.
by llm_trw 2y ago
Not to arbitrary depth. Empirically the majority of text has depth less than 6 when represented as a tree. Proofs on the other hand have no a priori limits to their depth.
671.
by llm_trw 2y ago
It's not so much that we should stop looking at problems because Kant said it's impossible. Rather that we should read what previous generations thought about this because for the first time in history we have a practical way to s
672.
by llm_trw 2y ago
Transformers are fundamentally linear, proofs are either tree like or dag like. You'd need a new attention mechanism to allow hierarchical reasoning.
673.
by llm_trw 2y ago
I wouldn't. The majority of problems there have mathematical optimizations which real problems never do. Worse if you start thinking in the way those problems encourage you to your programs will be completely invalidated even with a ti
674.
by llm_trw 2y ago
Again if it's so great fork the kernel and make it great. People will flock to you if you can manage.
675.
by llm_trw 2y ago
The flashlight is the radio tower, the tree is the tree, and the radio in the car is your eyes. There is no analogy here, it is literally the same EM waves shifted up to where our eyes can see them. It's like saying that the violins is
676.
by llm_trw 2y ago
Give it another month and it will be porn, just like sdxl.
677.
by llm_trw 2y ago
https://imgur.com/ayAWSKr Prompt: frozen cyan peach fuzz, with default settings on a first generation SD model. People _seriously_ do not understand how good these tools have been for nearly two years already.
678.
by llm_trw 2y ago
If you care enough you can email me at omni_vision_ai@proton.me I'd be happy to talk more in a less public setting.
679.
by llm_trw 2y ago
Markdown is a very poor language to try and use for anything other than single column typewriter like text. As evidenced by the fact that every project which uses it for more than that adds arbitrary extensions. The minimum viable language
680.
by llm_trw 2y ago
>Prove how useful LLMs are by doing something useful or value-generating with them. Then people will believe you. Talk is cheap. You after chat GPT2 was released.
681.
by llm_trw 2y ago
>Could you elaborate please ? No. I've tried explaining this to supposedly smart people in both a 15 minute pitch deck and a research paper and unless they were inclined to think it from the start no amount of proof has managed to c
682.
by llm_trw 2y ago
You could use messenger pigeons if you felt like it. People really don't understand how much better LLM swarms get with more agents. I never hit a point of diminishing returns on text quality over two days of running a swarm of llama2
683.
by llm_trw 2y ago
Because when you see someone try and reinvent Erlang in another language for the Nth time you know you can safely ignore them.
684.
by llm_trw 2y ago
An AI conference that isn't bullshit hype? Will wonders never cease? > Nobody presented multi agent simulations but I agree with you that is a very interesting way of thinking about things. To answer your question I did build a simu
685.
by llm_trw 2y ago
There isn't one. The real challenge for at scale inference is that the compute for models is too long to keep normal API connections open and you need a message passing system in place. This system also needs to be able to deliver larg
686.
by llm_trw 2y ago
At 1200 baud, yes.
687.
by llm_trw 2y ago
You would imagine wrong. The internet had settled just about all the major protocol decision by 1984 it was all about the hardware you could connect to the network.
688.
by llm_trw 2y ago
https://en.wikipedia.org/wiki/Acoustic_coupler >Prior to its breakup in 1984, Bell System's legal monopoly over telephony in the United States allowed the company to impose strict rules on how consumers could a
689.
by llm_trw 2y ago
>No, I don't trust Google neither other monopolies. If it's any conciliation neither does the US government. We are in for a ma-bel shakeup of the market. You know, the one that let internet running to your house happen.
690.
by llm_trw 2y ago
The 2018 server in my garage has had a better uptime than aws in the last 6 years.

More ›