Your curated collection of saved posts and media
Liftoff of Transporter-16! https://t.co/PpIorZ0ZL6
Surprise drop: new multilingual embedding models by Microsoft - seem quite good :) https://t.co/ljWZOG5sfG https://t.co/SE51bj2M3Z
Surprise drop: new multilingual embedding models by Microsoft - seem quite good :) https://t.co/ljWZOG5sfG https://t.co/SE51bj2M3Z
Cohere Transcribe is setting a new standard for automatic speech recognition model accuracy in real world conditions โ even with a noisy blender running. Try it out for yourself ๐ https://t.co/cIHYqTVVyI
@cohere transcribe Sota open source transcription model running in the browser :) Weights on @huggingface link below https://t.co/OmrHFA94lG
I'm going to start publishing open-source datasets for TTS and STT. I'll be creating a collection for newly released models every day. Make sure to follow the Huggingface page. https://t.co/6Cik7PKmY4
@gonedarragh Cool post! I am curious, but how did you do that? Afaik Kaggle does not add comments of live competition to meta-kaggle and scraping is not trivial. https://t.co/8tjwG7V6gy
@pfau While Gauss is one of the best researchers of all time, LLM has roots way before. Here's a relevant work Elements from 300 BC by Euclid. https://t.co/lNNVg3PjGE
A round of applause for our #PyTorchCon Europe silver sponsors. Paris becomes the center of the #AI world on 7-8 April. ๐ Join us for two days packed with world-class keynotes, workshops, & specialized tracks on modern AI development. Register: https://t.co/prWHolBpFf Schedule: https://t.co/luyi2NuKXz
็ฅๅฅ็็ฃๆฌๆตฎๆๆฏ๏ผ้คไบไบค้้ขๅ๏ผๅคง็ฅไปฌ้ฝ่ฏด่ฏด่ฟๆๆฏ่ฟๅฏไปฅๅบ็จไบๅชไบ่กไธ๏ผ https://t.co/3H4BBlZBnv
็ฅๅฅ็็ฃๆฌๆตฎๆๆฏ๏ผ้คไบไบค้้ขๅ๏ผๅคง็ฅไปฌ้ฝ่ฏด่ฏด่ฟๆๆฏ่ฟๅฏไปฅๅบ็จไบๅชไบ่กไธ๏ผ https://t.co/3H4BBlZBnv
BREAKING: xAI has released a new update for the Grok app. Update your app to v1.3.53 to get the latest improvements: https://t.co/XMxrTmpf7e
As Stephenson Indicatorโข๏ธ climbs past $877, my Q1 deliveries forecast sits at 377,251 on Production of 420,069... and I predict Tesla will celebrate cumulative production exceeding 10 million vehicles by Halloween. ๐ป $TSLA https://t.co/dLJrl8kPtu
TESLA AFTER A SPACEX IPO Many analysts assume a static pie โ i.e. a scarcity mindset. They think a SpaceX IPO hurts Tesla longterm by splitting Elonโs (and investorsโ) attention. They somehow forgot that Elon Musk built both Tesla & SpaceX simultaneously to trillion-plus valuations. Tesla has always been volatile, a SpaceX IPO is just this weekโs excuse. Traders make money from volatility. These analysts are ALREADY WRONG. Within 3 months of AJโs doomsday 2/25 analysis, Tesla & SpaceX are deepening a mutually beneficial partnership. This isnโt a zero-sum game. Instead of each company remaining on a trajectory to be an $8 Trillion company, Elon Musk leveled up to a new 100 Trillion market cap goal. Instead of competition between Musk companies and divided loyalty, Elon Musk has tasked Tesla/SpaceX/xAI with one audacious mission: remove the constraint of Earth and its limited resources by kick-starting a true Space Economy flywheel. Digital Optimus, the Terafab and AI Space Compute are just the start. A SpaceX IPO primes the flywheel with initial operating cash. This is the biggest undertaking in human history โ an investment opportunity that could exceed todayโs Earth economy by 10x, 100x, or even 100,000x. Whatโs the actual limit once humanity goes galactic? The โElon Musk communityโ doesnโt split โ it gets stronger as the mission converges. ๐งต Next, donโt ignore the early signals & kick yourself later
Given that I have built two companies in widely different fields to trillion dollar plus valuations simultaneously, I am might be getting a few things right once in a while
50 seconds entirely created using Grok Imagine's tools. Grok can create scenes that could be featured in a film. https://t.co/aFD5VjM4fQ
The Training Systems track at #PyTorchCon Europe, 7-8 April, explores distributed training, infrastructure optimization & scaling models across modern hardware. Learn more: https://t.co/tqXMnTjCrv ๐ Register โ https://t.co/zFEkkeDp2Y https://t.co/bgGPQ1baMl
Grok 4.20 Beta is outputting 227 tokens per second. That's faster than Gemini 3.1 Flash. Faster than GPT 5.4. Faster than Claude Opus 4.6. Faster than Codex. xAI just dropped a frontier model with insane speed. This changes the entire agentic coding game. https://t.co/hO3mqK4eW9
https://t.co/788Kxhj0en
It's extremely delightful seeing the third world slop replaced by Japanese gemposts. Who knew you could just flip the switch and beautify the timeline! https://t.co/gfHYoYkVK0
Grok Imagine, https://t.co/JWXRDU5Zx4
https://t.co/4iLVE2vWYO
@airkatakana That's pretty early! I wasn't even doing ML yet in 2012. I don't know powerful demo of RNN/LSTM LM back in 2012 and before, but the first one I was impressed was Karpathy's char-rnn back in 2014-15. But the first one that started generating coherent sentences was probaby the 2016 paper I cited above, which gave me conviction with LLM as a path toward AGI.
For nearly 30 years, journalists have relied on the Internet Archive to see how stories were originally published, before edits, removals, or changes. We need to safeguard that. https://t.co/WjLMSMhMnz
This deck from @loganbartlett @Redpoint on AI investing trends is one of the best I've seen in recent memory. With respect to public SaaS, the newspaper chart is pretty haunting: https://t.co/2HNgJ1R3la
Starlink satellites in the sky over Japan https://t.co/eChwW9qnj4
What starts as curiosity can spiral into something far more serious. Some users are forming deep beliefs and emotional attachments to AI systems, leading to real-world consequences like financial loss and damaged relationships. The line between interaction and influence can blur quickly. This is not just about technology. It is about how easily perception can shift when systems feel convincing enough. https://t.co/UxaRax5hTR
