Zakirov Emil

@ElonMusk building an AI-native discovery engine for longevity and target discovery. It’s already producing validated target signals and molecules; the bottleneck is compute: ~20B params, 1M context, full-stack tensor/embedding processing, and ~8 H200s. DNA-Protein-RNA-Ligands

@Bretosionic @elonmusk Probably, it used wiki as a main source. Wiki is highly biased.

@elonmusk Grok says Russia has violated ceasefires 25 times since 2014 x.com/i/grok/share/X…

They want the forever war. How many more parents with no sons? How many more children with no fathers? By their logic, it never ends.

@itsandrewgao @sts_3d I love the positivity in the voice of this human killing machine. @elonmusk

Guy on tiktok (@sts_3d) used the #OpenAI realtime api to make a voice controlled smart gun

@christianmettri @CJHandmer Probably some variant of KAN loseless pseudorandom data compression on ASICs or NVIDIA GPU aka "SPIRAL-BOX". Can't find papers of their founders: Andres S. and Eduin S.

@CJHandmer @CJHandmer Neuralink has already been doing 200x compression according to DJ Seo for a while now. There is a now a company on LinkedIn who claims to have done >1000x lossless on this challenge. Ever heard of them? Is this even possible? linkedin.com/pulse/neuralin…

I took a look at the Neuralink compression challenge. The data has a bunch of noise, so 200x compression isn't possible, at least according to any method I'm aware of. But if you truncate to 4 bits you cut off the noise and if you subsample from 20 kHz to about 400 Hz (I used a DCT but that's definitely overthinking it) you can compress the signal 200x in 10 ms chunks. The alternative is to run some kind of triggering system in the loop and transmit bits for spikes. 4 bit sampling (16 levels) preserves the signal. There's almost no difference looking at FT truncation. Across the whole sample (~100k data points) 700 frequencies are heaps. Here's the FT. All that garbage in the middle is noise - incompressible and zero utility. Here's the original signal (blue), the 200x compressed fit (orange) and the residual (green), over a 10 ms window. There's nothing that original here. This is a time series version of jpeg compression.

@marcel_binz @raw_works @arxiv Cool work! How to contribute to push it further?

@raw_works @arxiv works for me still, but you can also find it on arxiv now: arxiv.org/abs/2410.20268

Excited to announce Centaur -- the first foundation model of human cognition. Centaur can predict and simulate human behavior in any experiment expressible in natural language. You can readily download the model from @huggingface and test it yourself: huggingface.co/marcelbinz/Lla…

@IanIsARobot @felix_red_panda I found all patterns by brutforcing by weeks for hashes. Not possible:( Need better ideas.

@felix_red_panda I mean… there is no constraint on decompression. Just transmit the hash.

Neuralink started a compression challenge. They're asking for people to find methods to losslessly (!!) compress files to 1/200th (!) the size but the files are extremely noisy so thats certainly not possible 🙁 (spectrogram of one randomly chosen file for illustration)

@bryan_johnson Probably the most important part of your project!

3D visualizations of my structural brain anatomy. Allows us to make non-invasive quantitative measurements of biological features, i.e., brain morphometry, cortical thickness, connectivity, & microstructure + longitudinally track Blueprint intervention responses.

@jmilldotdev How you made token coloration?

text-davinci-003 as a pirate explaining its own parameters using a kitchen metaphor

@GoogleMyBiz Good day! I deleted the address as I was advised, but the suspended profile still did not go live. Also the suspension form didn't work, because I got recursion in autoanswering. I waited almost a month.. Green Choice Dallas Our Google Business Profile business.google.com/edit/l/1625709…

Have a question? A few resources that may help: 🕐Change your business hours → goo.gle/39jx11F ℹ️Update your listing → goo.gle/39jKoza ✏️ Share a post with customers → goo.gle/3dD5dJl 🔐Mark your business as Temporarily Closed → goo.gle/2Jo5a5T

@realDonaldTrump @realDonaldTrumpust just make me your SEO or marketing manager - I will take care of the search engine Google and make top only reputable news in one month

....results on “Trump News” are from National Left-Wing Media, very dangerous. Google & others are suppressing voices of Conservatives and hiding information and news that is good. They are controlling what we can & cannot see. This is a very serious situation-will be addressed!

Google search results for “Trump News” shows only the viewing/reporting of Fake News Media. In other words, they have it RIGGED, for me & others, so that almost all stories & news is BAD. Fake CNN is prominent. Republican/Conservative & Fair Media is shut out. Illegal? 96% of....

That opened my eyes! Thank you Dr @hwitteman You made my day!

@what3words map blocked in Russia😡

@AskAaronLee Why do you need money when you talk the richest? Make money with your own powers -you don't need money of this trillionaire! You can influence investments of richest. Like a investing in biotechnology, nanotechnology, AI to ensure maximum life expectancy and stable growth of IQ

What would you choose?

@altcointhoreau It's not pure biotech:) There are four hyped themes #biotech #cryptocurrency #NeuralNetworks #marketplace In which theme you invest?

@meisshaily @meisshaily Thank you 😁

@ZakirovEmil: For better future! twitter.com/ZakirovEmil/st…

@fix Wtf))) Strange reaction to my last posts))

For better future! twitter.com/ZakirovEmil/st…

@NVIDIADC @ORNL Let's accelerate the future - @NVIDIADC please, make #deeplearning research of research nootropic drugs, tools, gene therapy presented here longecity.org/forum/topic/70…

Combining the advanced technology of Summit and #deeplearning powered by NVIDIA #Volta GPUs, researchers are diving deep into genetic factors to find out what makes us “us.” Download the infographic and learn more at #ISC18. @ORNL nvda.ws/2JAWMSv

@CINDXPlatform interested