While deep learning is often shrouded in hype disconnected from reality, there are important first-principles reasons why deep learning is awesome. A thread... https://t.co/2ZDbCYd6MQ
— François Chollet (@fchollet) Feb 27, 2023
Today we're publicly releasing LLaMA, a state-of-the-art foundational LLM, as part of our ongoing commitment to open science, transparency and democratized access to new research. Learn more & request access ➡️ https://t.co/8AeLVhMWkq https://t.co/1BEkTngtnM
— Meta AI (@MetaAI) Feb 24, 2023
If you're interested in fine-tuning Stable Diffusion, check out Amazon SageMaker JumpStart's platform, which makes it easy to do so: https://t.co/eywOUPpFO5
— Stable Diffusion (@StableDiffusion) Feb 23, 2023
Pluggdin: Short AI and tech news app https://t.co/Xv4tH9oMDe
— Product Hunt Feed (@phfeed) Feb 23, 2023
Between 2019 and 2022, Spotify $SPOT spent more than a billion dollars on acquisitions related to podcasting, across both content and advertising/analytics technology. And early this year, the company abruptly changed its strategy. (1/X) https://t.co/7Wx0oBfoEr
— Eric Seufert (@eric_seufert) Feb 22, 2023
Imagine believing in an idea, but everyone says it will fail. Nobody will fund you. And the industry is experiencing a nuclear winter. 👇
— a16z (@a16z) Feb 22, 2023
@xamat Worth comparing to https://t.co/yKNCnUaNFq
— Emad (@EMostaque) Feb 22, 2023
Microsoft Research has released BioGPT, a large language model trained on biomedical research literature. The model achieves better-than-human performance on answering questions from the biomedical literature, as evaluated on PubMedQA. The code for the https://t.co/MEPL0wmjFo… https://t.co/Bctt1GLrdO https://t.co/HPa7g5f2iy
— Bojan Tunguz (@tunguz) Feb 21, 2023
2-21-2023 Update •Section 230 in the Supreme Court •Reach and Speech •The First Amendment and U.S. Speech Controls https://t.co/hvpkP1d16Z
— Stratechery (@stratechery) Feb 21, 2023
People + AI >> AI alone "The tactics that put a human back on top on the Go board were suggested by a computer program that had probed the AI systems looking for weaknesses. The suggested plan was then ruthlessly delivered by Pelrine." https://t.co/z5PXxf20q8
— Vijay Pande (@vijaypande) Feb 20, 2023
Writing an engineering strategy https://t.co/Uo0OLlzOgA (https://t.co/9A9kdYOZK8)
— Hacker News 200 (@newsyc200) Feb 20, 2023
"Commercial casinos in the United States won more than $60 billion from gamblers in 2022... tribal casinos could report an additional $41 billion in revenue later this year, putting the total U.S. casino industry over the $100 billion mark." https://t.co/oIGKBPzbX1
— Tren Griffin (@trengriffin) Feb 20, 2023
https://t.co/9LeXJQUAuh
— Gopi Vikranth (@GopiVikranth) Feb 20, 2023
6/ "GPT is all you need for the backend" https://t.co/Wu7XOqFHbi Tired: use an LLM to help you write a backend Wired: LLM is the backend Inspiring project from a recent Scale hackathon. The LLM backend takes state as JSON blob and modifies it based on... English description. https://t.co/k4So1luWkX
— Andrej Karpathy (@karpathy) Feb 19, 2023
Think @elonmusk basically ended the old world of verification and badges online forever. https://t.co/vOTDrB8f8v
— Sriram Krishnan - sriramk.eth (@sriramk) Feb 19, 2023
More info here. One differentiator between this and Twitter Blue is that you need to verify your identity with an actual government ID. But a lot of similarities for sure. https://t.co/AqbSgI0O1X
— Kurt Wagner (@KurtWagner8) Feb 19, 2023
To those that have read #BUILD and are looking for a little extra magic - give General Magic a watch 💫 https://t.co/y4BYDFxJp0
— Tony Fadell (@tfadell) Feb 18, 2023
Have a listen for yourself. https://t.co/3qVPlS3L8h
— Futurism (@futurism) Feb 17, 2023
Don't just prototype the product, prototype the experience.
— Tony Fadell (@tfadell) Feb 17, 2023
Scoop: We got our hands on Stripe's fundraising deck. Here's a look at how the company is trying to convince investors to overlook a sharp slowdown in revenue last year. https://t.co/cqglcRhkwK w/ @coryweinberg
— Kate Clark (@KateClarkTweets) Feb 17, 2023
https://t.co/ZGoa2czu8G
— Sandeep | Polygon 💜 Top 3 by impact (@sandeepnailwal) Feb 17, 2023
All the tweets above were published in January. (Before the announcement of Bing Chat aka Sydney) Here is the latest thread with Yann LeCun's opinion about LMMs. https://t.co/VaG2EyzHtI 6/6
— TheSequence (@TheSequenceAI) Feb 17, 2023
Research 1 Researchers from @MIT, @GoogleAI, and @Stanford published a paper discovering an algorithm that can describe the in-context learning of LLMs. Go check it out https://t.co/uQT773jy9u https://t.co/dFPZDUc7RY
— TheSequence (@TheSequenceAI) Feb 16, 2023
The goal of consumer tech development used to be pretty simple: design and build something of value to people. But now *we* are the product, not the prime beneficiary. https://t.co/rPhgcidsDF
— MIT Technology Review (@techreview) Feb 16, 2023
From Bing to Sydney https://t.co/jTD2fsOSWg (https://t.co/2N5FWM2jA2)
— Hacker News 200 (@newsyc200) Feb 15, 2023
The Adam optimizer is at the heart of modern AI. Researchers have been trying to dethrone Adam for years. How about we ask a machine to do a better job? @GoogleAI uses evolution to discover a simpler & efficient algorithm with remarkable features. It’s just 8 lines of code: 🧵 https://t.co/a4A03Z8egs
— Jim Fan (@DrJimFan) Feb 15, 2023
Loyal to the Company https://t.co/Sm7GRK0WtG (https://t.co/T4VABs541J)
— Hacker News 200 (@newsyc200) Feb 14, 2023
New research from Meta AI reduces latency of existing Vision Transformer models with no additional training. ToMe combines similar tokens, reducing computation w/o losing information. Results are 2-3x speed for state-of-the-art models w/ minimal performance loss. Read more ⬇️
— Meta AI (@MetaAI) Feb 13, 2023
Toolformer: Language Models Can Teach Themselves to Use Tools - paper by Meta AI Research #nlp #ai https://t.co/rOfo8BaHsU
— /MachineLearning (@slashML) Feb 12, 2023
The architecture of GPT3 (napkin size) https://t.co/x6sSMpIvmG https://t.co/keUqFAr9CT
— Grady Booch (@Grady_Booch) Feb 12, 2023
Net Income, 2021 (Billions)... Apple: $101 Microsoft: $71 Google: $76 Amazon: $33 Facebook: $39 Net Income, 2022 (Billions)... Apple: $95 Microsoft: $67 Google: $60 Amazon: -$3 Facebook: $23 $AAPL $MSFT $GOOGL $AMZN $META
— Charlie Bilello (@charliebilello) Feb 12, 2023
Great AMA with Meta's CTO on 2/2/23. Highlights: 1. Recent events highlighted VR "isn't mass market enough" (v. MR/AR). Mammals need to see real world 2. AR breakthroughs needed to achieve critical mass include neural interface 3. AR tradeoffs are "a https://t.co/j6rIIxTFHn… https://t.co/naNIs07auJ
— Matthew Ball (@ballmatthew) Feb 12, 2023
Hard Prompts Made Easy (PEZ) Want to generate a text prompt for your image that is useful for Stable Diffusion? try out the @Gradio demo on @huggingface by @ywen99 : https://t.co/PiOBLSqWGP https://t.co/PzlmHg8Kd3
— AK (@_akhaliq) Feb 9, 2023
"What we want to do is make a leapfrog product that is WAY smarter than any mobile device has ever been, and super-easy to use. THIS is what iPhone is." — Steve Jobs, MacWorld, 2007 https://t.co/d2sLFcFzUU
— Steve Jobs Stories (@SJobs_Stories) Feb 7, 2023
The first-ever American Dynamism Summit brought together industry experts and government officials who are determined to solve tough problems and build a better future. They shared their thoughts in this series of videos 🧵👇 https://t.co/BQaQjPA92W
— a16z (@a16z) Feb 7, 2023
Zero-shot Image-to-Image Translation Proposes pix2pix-zero, an image-to-image translation method that can preserve the content of the original image without manual prompting. proj: https://t.co/QBjTVa5V9e abs: https://t.co/bnvSFFa99v https://t.co/Ub7eOS6KBK
— Aran Komatsuzaki (@arankomatsuzaki) Feb 7, 2023
1/ In 2021, we shared next-gen language + conversation capabilities powered by our Language Model for Dialogue Applications (LaMDA). Coming soon: Bard, a new experimental conversational #GoogleAI service powered by LaMDA. https://t.co/cYo6iYdmQ1
— Sundar Pichai (@sundarpichai) Feb 6, 2023
The Four Horsemen of the Tech Recession Tech is increasingly divorced from the real economy thanks to the COVID hangover and Apple's App Tracking Transparency https://t.co/Hz49aSbXMW
— Stratechery (@stratechery) Feb 6, 2023
A breakthrough every single week! Video Diffusion Models are General Video Editors: https://t.co/R90SVDAha9
— Suhail (@Suhail) Feb 5, 2023
Google announces Dreamix: a model that generates videos when given a prompt and an input image/video. #dreamix #google https://t.co/k1vf22t1di
— /MachineLearning (@slashML) Feb 4, 2023
Bill Gurley: "Matt Ridley talks about ideas having sex and how that can impact innovation and increase productivity." https://t.co/0jgHolT62Z https://t.co/s5x900Qx86
— Tren Griffin (@trengriffin) Feb 4, 2023
Price to Earnings Ratios... Amazon: N/A (neg TTM EPS) Tesla: 52x Netflix: 37x Microsoft: 29x Apple: 26x S&P 500: 24x Google: 23x Facebook: 22x Price to Sales Ratios... Microsoft: 9.5x Tesla: 8.1x Apple: 6.4x Netflix: 5.2x Google: 4.9x Facebook: 4.3x S&P 500: 2.4x Amazon: 2.1x
— Charlie Bilello (@charliebilello) Feb 4, 2023
Google announces Dreamix: a model that generates videos when given: - video + prompt (Video editing) - input images + prompt (Subject Driven Generation) - input image + prompt (Image-toVideo https://t.co/Ht3HKjAvet
— bleedingedge.ai (@bleedingedgeai) Feb 3, 2023
The most practical open-source competitor to @OpenAI 's GPT-3 is Google's Flan-T5 Here are 5 Flan-T5 resources to try out easily, deploy, or fine-tune it! 🧵
— Ramsri Goutham Golla (@ramsri_goutham) Feb 3, 2023
I wrote about Adani's backstory in September (his wealth his since fallen from $140B to $70B). For a summary of the Hindenburg-Adani battle, I've updated that article with happenings over the past week (which has seen Adani's conglomerate lose $100B+). https://t.co/ZntCRR8esx
— Trung Phan (@TrungTPhan) Feb 3, 2023
Vector embeddings have become more popular with @OpenAI's reduced embedding pricing and the release of chatGPT! But how did we evolve from word -> sentence -> document vectors? Here is my video covering the same! https://t.co/KhGjkvP1v6
— Ramsri Goutham Golla (@ramsri_goutham) Feb 2, 2023
If you could only have 1 book in your bookcase, which one would it be? https://t.co/N8pXPduftG
— Compounding Quality (@QCompounding) Feb 1, 2023
Podcast with the brilliant @pmarca is out! We discuss: - how AI will revolutionize software - are NFTs useless, & should he fund flying cars instead - a16z's biggest vulnerability - future of fusion, education, Twitter, venture, managerialism, & big tech Interesting throughout! https://t.co/B3vQ5FSix0
— Dwarkesh Patel (@dwarkesh_sp) Feb 1, 2023
Why is making a dark mode greyscale so hard to get right? Well, of course it has to do with the weird way humans perceive colour and contrast. 👇 https://t.co/3Z8SKxghSk
— Dan Hollick 🇿🇦 (@DanHollick) Feb 1, 2023
"We’re launching a classifier trained to distinguish between AI-written and human-written text." https://t.co/nxsPUDwzqM
— Adam.GPT (@TheRealAdamG) Jan 31, 2023
Find Datasets for your Next Machine learning Project: Kaggle → https://t.co/FPYC3aKdtB Papers with Code → https://t.co/aAWnRKjJg8 Github → https://t.co/MWBfAs7QtE Reddit → https://t.co/mZXVwhC9Eh HuggingFace → https://t.co/EXHy4A61AT Open ML → https://t.co/id3EARaIUU
— Sumanth (@Sumanth_077) Jan 31, 2023
