Really enjoyed this chat with Sonny! Discussed https://t.co/efQJ4oo2Rc, churn, what's to come in SVOD, and IP. https://t.co/wnBiBDfsye https://t.co/YZBDoefvRl
— Matthew Ball (@ballmatthew) Apr 30, 2023
https://t.co/VG8pNfj877
— Gopi Vikranth (@GopiVikranth) May 1, 2023
Step 1: Install Chrome Canary Step 2: Run Vicuna-7B LLM right on your browser. More info: https://t.co/Z5ytu0GnTJ Repo: https://t.co/aofLSJSNrO https://t.co/b4wyYfpBI5
— Mahesh Sathiamoorthy (@madiator) Apr 30, 2023
Sharing my latest article - "Berkshire Hathaway - The World’s Greatest Serial Acquirer of Businesses". $BRK What became apparent was that Buffett and Munger with Berkshire were not just great investors, but also great serial acquirers. Sharing the flywheel of Berkshire. Enjoy! https://t.co/dvwlWs2bBr
— Eugene Ng (@EugeneNg_VCap) Apr 30, 2023
Let this sink in: AI now performs better at Theory of Mind tests than actual humans. GPT-4 scores 100% on 2-shot chain-of-thought and step-by-step reasoning benchmarks. (Human accuracy on ToM tests is ~87%) https://t.co/Pr5MBKTBG7
— AI Breakfast (@AiBreakfast) Apr 30, 2023
My full talk on evolving conversational interfaces forward with LLMs from @mlopscommunity is live! 🎉 https://t.co/iSpoP5vEhR https://t.co/uTmTjJ5Pdd
— Linus (@thesephist) Apr 29, 2023
New AI just dropped! Runs offline on your iPhone - you can install it from TestFlight https://t.co/kmuNFILVLh
— Simon Willison (@simonw) Apr 29, 2023
A new text-to-image model is here! DeepFloyd IF by folks at @StabilityAI. And, wait for it ... it's open source! :) Try it out locally (you'll need ~14 GB of VRAM on your GPU) or here on HF Spaces: https://t.co/8ASPeNWTdg Code: https://t.co/IEIvDP68nA 1/ https://t.co/PQo2MlG8WN
— Aleksa Gordić 🍿🤖 (@gordic_aleksa) Apr 29, 2023
Open Source innovation in Generative AI has been nothing but frenetic, and the latest edition of GenOS Index reflects that: 18 new entrants with a median age of only 41 days, and the topmost project, AutoGPT, racking up 117k stars in 7 weeks! https://t.co/TruOguyS04 https://t.co/ErskfOgvGh
— Sudip Chakrabarti (@chakrabartis) Apr 28, 2023
🚀 LLMs have got a new glow 👉Create complex chat apps in just 4 clicks with our 🦋ChainFury implemented using powerful @LangChainAI framework by @hwchase17 . Try now: https://t.co/2WoU3OfCIM More Features: thread 🧵 https://t.co/Pi6hH40oXb
— Priya S (@PriyaSridhara13) Apr 28, 2023
Head over to https://t.co/hryfpNRa0S to see various links and resources to LLMs (some sections seem to be still under construction). Curating this graph should have taken a lot of time :) https://t.co/IUoUp1G60X
— Mahesh Sathiamoorthy (@madiator) Apr 28, 2023
If you’re using @LangChainAI and working with really messy or unstructured CSV/Excel/GSheet data, here are some tips if you’re using Python:
— matt turk (@TurkMatthew) Apr 27, 2023
Thanks to parameter-efficient finetuning techniques, you can finetune a 7B LLM on a single GPU in 1-2 h using techniques like low-rank adaptation (LoRA). Just wrote a new article explaining how LoRA works & how to finetune a pretrained LLM like LLaMA: https://t.co/zxi4gNcJdX
— Sebastian Raschka (@rasbt) Apr 26, 2023
The music business is about to change forever now that AI can make Drake's songs better than Drake can. And Grimes has a brilliant proposal for the future of AI music. This isn't the next Napster moment. It's bigger than that. Here is what you need to know: https://t.co/uAwQXPszEB
— Nathan Lands (@NathanLands) Apr 26, 2023
2. Runway launches Gen-1 iPhone app Runway introduced its first mobile app for iOS, allowing users to access their Gen-1 video-to-video generative AI model directly from their phones. Soon, people will be creating entire films right from their pockets! https://t.co/YFF9DXDbSr
— Rowan Cheung (@rowancheung) Apr 26, 2023
Meet Repo Chat! It performs Q&A on your code repos. How it works: - ingest code via GitHub url - embed files w/ OpenAI embeddings - create vectordb w/ Supabase - perform document retrieval + q&a powered by Langchain - input via Python CLI GitHub: https://t.co/JHNehZkIt1 https://t.co/bsP0XLlOA5
— Mckay Wrigley (@mckaywrigley) Apr 25, 2023
Resources for deepening knowledge of Transformers #transformers #nlp https://t.co/Ff36oeu3W9
— /MachineLearning (@slashML) Apr 25, 2023
Sort of a side note, but just realized how ridiculously easy it is, with @LangChainAI doing the heavy lifting, to use #GPT-4 to build structured, evaluative info from basically any unstructured documents that can be loaded from URLs! Here's a quick demo: https://t.co/ypgrCIjybw https://t.co/M5K7OzEq6L
— Martin Zirulnik (@MartinZirulnik) Apr 24, 2023
Investing in index funds can make you a tax-free millionaire, using a Roth IRA. There are over 2,500 index funds to invest in, so here are my top 5 (to help you build wealth):
— Andrew Lokenauth (@FluentInFinance) Apr 24, 2023
How ChatGPT works technically. We attempted to explain how it works in this video. We will cover: • Large Language Model • GPT model • Fine-tuning • Prompt engineering • How to answer a prompt Watch and subscribe here (YouTube video): https://t.co/Kv0dqSPajX https://t.co/YPsJYtCFzH
— Alex Xu (@alexxubyte) Apr 24, 2023
Emad Mostaque claimed to investors that Stable Diffusion was made by Stability AI. Not true (as I never tire of explaining whenever this myth appears): SD was a German research project at the University of Heidelberg (now Munich). Stability came later. https://t.co/wGZakidLxp
— Silke Hahn (@_SilkeHahn) Apr 24, 2023
What are the three pillars of 𝗥𝗲𝗮𝗹 𝗧𝗶𝗺𝗲 𝗠𝗼𝗱𝗲𝗹 𝗢𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆? Your Machine Learning Systems can fail in more ways than regular software. Machine Learning Observability System is where you will continuously monitor ML Application Health and Performance.… https://t.co/Lqsgm6Bj8E https://t.co/TI6HRkxiE8
— Aurimas Griciūnas (@Aurimas_Gr) Apr 24, 2023
"This is how GPT-4 sees and hears consciousness" I used GPT-4 to describe consciousness. Then I used its description to generate an image, a video based on this image, and a soundtrack. Tools I used: GPT-4, Midjourney, Kaiber AI, Mubert, RunwayML This is the… https://t.co/qDROCZKrTy https://t.co/iCgCNGtCam
— Kris Kashtanova (@icreatelife) Apr 24, 2023
The Little Book of Deep Learning A very concise/brief book on deep learning. Covers almost any topic you'd want to know today from foundations, efficient computation, model architectures, training models, synthesis(generative AI), etc... And it's free: https://t.co/HyKmcFJWSx https://t.co/1QMrwblxix
— Jean de Nyandwi (@Jeande_d) Apr 23, 2023
Midjourney - Styles: part 1 Here are 15 different illustration & design styles you can use with Midjourney to give your images a special touch. They can be easily integrated into basic prompts or more complex scenes. 🧵Thread with visual examples 👇 https://t.co/qIljNSe5uM
— Anonymouse (@TheMouseCrypto) Apr 23, 2023
I highly recommend you set aside 30 minutes to watch the @TEDTalks with @gdb, @OpenAI co-founder and president. https://t.co/aY7G35WTfP
— Paul Roetzer (@paulroetzer) Apr 22, 2023
This is a really cool paper: differentiable kernel sizes! https://t.co/8rmmt0Y8dK Being used in this work to perhaps replace Transformers with convolutions: https://t.co/HRpkR6fYdF
— Brad Neuberg (@bradneuberg) Apr 22, 2023
The ChatGPT we all wanted is finally here. With this new AI tool, anyone can use GPT-4 for free without a ChatGPT+ subscription. Plus, they have image generation, custom personas, a shareable chat, and more👇 https://t.co/dFY4spP1gt
— Rowan Cheung (@rowancheung) Apr 21, 2023
Excited to introduce NexusGPT, a marketplace of 800+ autonomous AI-agent with each specific skillset. Demo is free on https://t.co/diCTFrBxrR had great inspiration from @dzhng's BabyAGI built on top of @LangChainAI using @OpenAI and @trychroma to keep conversation memory https://t.co/3ZTbifsLev
— achammah (@achammah1) Apr 21, 2023
Today we’re updating Bard with the ability to help people with programming and software development tasks. We’re launching these capabilities in 20+ programming languages including C++, Go, Java, Javascript, Python and Typescript. https://t.co/MFDVPzV5O6 1/
— Jack Krawczyk (@JackK) Apr 21, 2023
Runways Gen2 text-to-video is here, and it's mind blowing. Truly another 0-to-1 AI moment. Here are some of the coolest examples of a new tool that just started rolling out in beta a couple hours ago 🧵👇 https://t.co/5phh5lQpli
— Nathaniel Whittemore (@nlw) Apr 21, 2023
There's a chance that LoRA finetunes work so well that it dramatically alters the finetuning vs. retrieval + few-shot prompting power dynamic in favor of the former for many applications. PEFT (Parameter Efficient Finetuning, LoRA included) are emerging techniques that make it… https://t.co/uPXSJJ684P
— Andrej Karpathy (@karpathy) Apr 20, 2023
Like a visual @LangChainAI I started building many internal tools like this for the team at my ecommerce company about a year ago. Then I decided to make a no-code version so I can whip 'em up even faster. Now anyone can make them with https://t.co/DbNcbuJlbj https://t.co/D98fxusmNx
— Rob Martin (@lasik82) Apr 20, 2023
lots of people have asked how to use langchain with self-hosted LLMs really excited to see this tutorial! @raydistributed are extremely knowledgable, anything they share is worth paying attention to https://t.co/xdVzGk9jr2
— Harrison Chase (@hwchase17) Apr 19, 2023
LlamaAcademy: a factory that teaches LLaMA's how write API code. ChatGPT Plugins, but no manifest file is needed -- it just reads the API docs like a human would! For example, here is a finetuned model that learned how to talk to Notion by reading its API docs: https://t.co/DPJQaUH3u5
— Daniel Gross (@danielgross) Apr 19, 2023
🤖LangChain x AutoGPT🤖 First BabyAGI, then CAMEL, then generative agents, now finally AutoGPT... Implementing the core AutoGPT prompt + agent executor in LangChain primitives (to allow for swapping between models, vectorstores, tools, etc) Notebook: https://t.co/2hag1gISd8 https://t.co/v8R3pht4ne
— langchain (@LangChainAI) Apr 19, 2023
I've seen 1,000,000 LLM demos and "AI influencer" posts. But none of them showed me how to actually deploy a real app. So, I built a project (open source) to figure it out, and it only uses 1 CPU and 1 GB RAM. Here's how I built a simple ChatGPT streaming backend in FastAPI:
— Mark Tenenholtz (@marktenenholtz) Apr 19, 2023
Another truly open-source instruction-following LLM was recently released - Dolly 2.0 by @databricks! I wrote a lot about LLaMA over the previous period, but there is a BIG caveat with LLaMA: the license is not really permissive... Blog: https://t.co/1M7HSbWgqk 1/ 👇🧵 https://t.co/n6or2WQxzl
— Aleksa Gordić 🍿🤖 (@gordic_aleksa) Apr 18, 2023
Last month was particularly exciting for open-source AI, featuring several open source implementations of large language models (LLMs). Now we are seeing a big wave of open-source datasets! We had Databricks Dolly 15k and OpenAssistant Conversations for instruction finetuning.… https://t.co/V9TTyxd5X0 https://t.co/gkSt0QKWNH
— Sebastian Raschka (@rasbt) Apr 18, 2023
Highly suggest reading this report yourself: https://t.co/KFu74r22hp
— Pete (@nonmayorpete) Apr 17, 2023
Following the release of LLaMA, we saw a rapid explosion of open-source research on large language models (LLMs). Here are the three most notable model releases during this time… 🧵 [1/8] https://t.co/OLaYzAqJhJ
— Cameron R. Wolfe (@cwolferesearch) Apr 17, 2023
4. Automatic Chain of Thought Prompting in Large Language Models Proposes an Auto-CoT method automating the Manual-CoT while achieving comparable results. Auto-CoT allows making CoT prompting more scalable and less talent-dependent. https://t.co/8BGjRNPPi1 https://t.co/i71fl34Pcd
— TuringPost (@TheTuringPost) Apr 17, 2023
AI Canon Top 3 (non fiction): 1. The Scaling Hypothesis @gwern https://t.co/KfpP7d6Rkb 2. The AI Revolution @waitbutwhy https://t.co/aqpw8gUYTu 3. What is ChatGPT doing? @stephen_wolfram https://t.co/I4IvC1Pkiw
— Siqi Chen (@blader) Apr 17, 2023
Let's see how the chaos parameter affects 4 images in the same grid by using a simple prompt: A woman looking at the camera Since the default chaos parameter is already set to 0, we should expect 4 similar images. Indeed, that's the case. https://t.co/CEPFVZoPFB
— Anonymouse (@TheMouseCrypto) Apr 16, 2023
https://t.co/VOVbpwr98j
— Gopi Vikranth (@GopiVikranth) Apr 17, 2023
Web LLM is insane. 1) Go download the latest Chrome beta, which shipped WebGPU support: https://t.co/XboDnsOLZN 2) Now use a 7B-param LLM in your browser! https://t.co/bFnW2fK5DR 3) Marvel at the "How" section on their GitHub: https://t.co/YNLMJiyJVk https://t.co/nQ7FcKHv8h
— Sergey Karayev (@sergeykarayev) Apr 16, 2023
I read all 154 pages of Microsoft Research's "Sparks of AGI". Most people don't realise the shocking reality AI is going to bring. Here's what I found:
— Alex Banks (@thealexbanks) Apr 16, 2023
Some of you asked me to make an A.I.-assisted comic book in Midjourney (after I shared one I made in Adobe Firefly) and share my process with you. "Different Input" is a sci-fi comic book I've been working on. And here's a tutorial of my process: https://t.co/6XMt2LhHfx https://t.co/8KVgOxtemI
— Kris Kashtanova (@icreatelife) Apr 15, 2023
The best prompts I've ever created. The thread will be continuously updated. Prompts are in ALTs or threads. https://t.co/uOjjvbKDZs
— Tatiana Tsiguleva (@ciguleva) Apr 15, 2023
Reading through the Segment Anywhere Model (SAM) paper, some notes & insights as I do so in this Twitter thread. This whole paper is quite incredible actually, very strong work 1/n https://t.co/HhJUoBakBc
— Brad Neuberg (@bradneuberg) Apr 15, 2023
The slides for my talk "Everything you wanted to know about ChatGPT (Except for what OpenAI doesn't tell us)" are available here: https://t.co/egZnJSjR2e. I used it for both a guest lecture in an applied ML class and for a talk for non-CS audience. https://t.co/mM83V6R43X
— Vered Shwartz (@VeredShwartz) Apr 14, 2023
If you're only using ChatGPT, you're missing out. Here is a SUPER simple way to start using OpenAI APIs (with screenshots and everything!) for all of my non-technical followers that want to use more of this tech.
— Allie K. Miller (@alliekmiller) Apr 14, 2023
Article in @QuantaMagazine on two of the most brilliant thinkers I know: Bruno Olshausen and Pentti Kanerva: https://t.co/hQ1fpWhyEm
— Melanie Mitchell (@MelMitchell1) Apr 14, 2023
⚡ From the authors of Microsoft DeepSpeed - DeepSpeed-Chat! A new library that allows you to easily train 10B+ ChatGPT-style models on a single GPU! Single script, RLHF support, and 15x faster than SOTA apparently! GitHub: https://t.co/2oe2F0OFeu 1/ 🧵👇 https://t.co/poyD49QOfo
— Aleksa Gordić 🍿🤖 (@gordic_aleksa) Apr 14, 2023
🧵How can we teach LLMs to reason? 🧵 “Chain of thought prompting can improve performance on various reasoning tasks... the benefits of chain of thought prompting only materialize with a sufficient number of model parameters (around 100B).” 🔗: https://t.co/Z5pSvNCADK [1/4]
— Cameron R. Wolfe (@cwolferesearch) Apr 13, 2023
🌵LangChain x GPTCache🌵 Caching can be useful in certain situations to save money on repeated LLM calls GPTCache is a project that caches requests/responses not only the exact prompt, but based on semantic similarity. h/t Frank Liu for adding Docs: https://t.co/pm1ubEldYV
— langchain (@LangChainAI) Apr 13, 2023
Here are some incredible things people are already doing with AutoGPT It feels like yesterday I discovered it... 🧵 A thread https://t.co/TwSMIpJGI3
— Aadit Sheth | Mr. Prompts (@aaditsh) Apr 13, 2023
@habits @threadreaderapp Unrolled, with a few extra takeaways. https://t.co/jsgiUXdDUO
— Conor Mac (@InvestmentTalkk) Apr 13, 2023
New post: bringing LLM applications to production! 1. Challenges of LLM engineering & the solutions that I’ve seen 2. How to compose multiple tasks and incorporate tools (e.g. SQL executor, bash, web browsers, third-party APIs) 3. Promising use cases https://t.co/XWmZfIc5mr
— Chip Huyen (@chipro) Apr 12, 2023
BREAKING OpenAI released a implementation of Consistency Models consistency models, a new family of generative models that achieve high sample quality without adversarial training. They support fast one-step generation by design, while still allowing for few-step sampling to… https://t.co/PUOLPUfDJv https://t.co/ZXIqtbexH0
— AK (@_akhaliq) Apr 12, 2023
Free Dolly! Introducing the first *commercially viable*, open source, instruction-following LLM. Dolly 2.0 is available for commercial applications without having to pay for API access or sharing data with 3rd parties. https://t.co/HtRVSmArFU
— Ali Ghodsi (@alighodsi) Apr 12, 2023
The Coming of Local LLMs https://t.co/D2moYzHjsa (https://t.co/vscnKa7Wze)
— Hacker News 200 (@newsyc200) Apr 12, 2023
"We went and bought the parts to build one hundred computers. We built fifty of them and delivered them. We got paid in cash and ran back and paid the people that sold us parts. ... our profit wasn’t liquid—it was in fifty computers sitting on the floor." https://t.co/uN3tvgNvFM
— Tren Griffin (@trengriffin) Apr 11, 2023
Autonomous AI is a new trend. 2 repos were highly talked about everywhere ▪️ Auto-GPT @SigGravitas ▪️ BabyAGI @yoheinakajima Original tweets, experiments, and opinions 🧵 https://t.co/COsXw04wUu
— TheTuringPost (@TheTuringPost) Apr 11, 2023
One of the main benefits of GPT-4 relative to prior models (like ChatGPT/GPT-3.5) is that the model is incredibly steerable. Here’s what this means and how you can use it to create better chat experiences… 🧵[1/8] https://t.co/rBB1WUorRx
— Cameron R. Wolfe (@cwolferesearch) Apr 10, 2023
My favorite people in AI to follow, the reason to follow them & their top tweet:
— Nathan Lands (@NathanLands) Apr 10, 2023
Stochastic parrots and glorified auto-completers are inductive processes, but how is it that a LLM like GPT-4 is able to perform abduction? https://t.co/bZ1c6lQBVK
— Carlos E. Perez (@IntuitMachine) Apr 10, 2023
5 latest open-source LLMs (save the list) 1. BLOOMZ 2. OPT-IML 3. Pythia 4. LLaMA 5. Vicuna All repos in 🧵 https://t.co/6fROyNSQQX
— TheTuringPost (@TheTuringPost) Apr 10, 2023
Super simple demo/proof of concept using @LangChainAI to give an AI agent access to the terminal. The agent opens the browser, searches for Lex Fridman on Youtube, and scrapes the contents of the page. All in less than 15 lines of code. https://t.co/CwCpPrYPKQ
— Tosh (@lobotomyrobot) Apr 9, 2023
https://t.co/qzd9GPbCih
— Gopi Vikranth (@GopiVikranth) Apr 10, 2023
An intriguing trend in AI 🤖: “Models all the way down” (aka "stacking") Have models invoke other models, then watch as emergent intelligence develops ✨ Here’s a discussion of what, how, and why this is important to watch 👇 https://t.co/hN3JND32CK
— Jay Hack (@mathemagic1an) Apr 9, 2023
Prefix tuning and adapters are the 2 out of the 3 most widely used parameter-efficient finetuning methods for large language models (LLMs). If you look closely, the recent LLaMA-Adapter method that made big waves is actually a prefix tuning method, not an adapter method. 1/2 https://t.co/Q9y0pn9NHS https://t.co/DN6TDfyNGA
— Sebastian Raschka (@rasbt) Apr 9, 2023
I love desserts, like a lot, maybe too much :D 💬 Midjourney Prompt and in the ALT: a slice of chocolate cake, professional color grading, soft shadows, clean sharp focus, high-end retouching, food magazine photography --s 250 https://t.co/Rg9G5ZdcY3
— Marcos (@followmarcos) Apr 9, 2023
GPT-4all with 20K starts in a very short period is a crazily successful project. In contrast, @huggingface has ~90K. GPT-4all: Demo, data, and code to train an assistant-style large language model with ~800k GPT-3.5-Turbo Generations based on LLaMa https://t.co/BVsBmZw60U
— Leo Boytsov (@srchvrs) Apr 9, 2023
Story telling as prompts! Inspired by @icreatelife @daidatep. Instead of descriptions, you can tell a story and let #midjourney interpret. The art of #synthography #AIart #MidjourneyV5 https://t.co/WgNWG0BEyQ
— Passio AI (@passio_ai) Apr 8, 2023
@icreatelife Before Kris's post I didn't know #AdobeFirefly can do this Prompt: Big Hollywood movie popculture franchise style infographics around a race car. Display car metrics. Next to it a running race car driver from the future. THANKS for the inspire Kris! #aiinforgraphics #infographics https://t.co/Oj5Wf67iyI
— Rik (@Rik_Vasquez) Apr 8, 2023
I should tweet more 😅. I made this reddit post and it got over 3 million views 🤯 honestly didn't think that many people would be seeing it. Got over 1000 new subs because of it 😭 Over 4000 subs now and loving every second of it 😈 https://t.co/7uTTTaQPIV
— Nofil (@Lifonnnn) Apr 8, 2023
Please, someone, help this guy 😄 #AdobeFirefly prompt by @ricarhartt : A glass jar containing a cute 3D character https://t.co/TCSB6RyWDS
— Kris Kashtanova (@icreatelife) Apr 7, 2023
RLHF might sound easy in theory, but in practice there are many things that can go wrong. A new post from hugging face shows how and why. They show how to perform RLHF on LLaMA https://t.co/qNyw957vCB https://t.co/wQtcATEN4i
— anton (@abacaj) Apr 6, 2023
This is an insanely cool GPT-4 experiment! Auto-GPT is an open-source application showing the power of LLMs like GPT-4 to autonomously develop and manage different kinds of tasks like completing a code session or suggesting a business idea. You give an agent an identity,… https://t.co/t9CttED78b https://t.co/TxtiG8WdyU
— elvis (@omarsar0) Apr 6, 2023
It is such a pleasure to read papers from Girshick's team (@inkynumbers). The Segment Anything paper is wonderful. https://t.co/ECgijKaMcc
— Thomas G. Dietterich (@tdietterich) Apr 6, 2023
A new function in Midjourney (pro plans in /fast mode) PERMUTATIONS. You can write parameters in { } and MJ will generate variations. This explainer is made in collaboration with @saana_ai Find more explainers: https://t.co/SRYyy9N5W9 https://t.co/c7u0zummMV
— Kris Kashtanova (@icreatelife) Apr 6, 2023
The new Permutations feature in Midjourney works so well with Additive Prompting Test out different camera angles, film types, clothing brands, models, locations, anything you want, by putting it in {brakets}. Then you can batch run up to 40 prompts in one go. This is what a… https://t.co/OMGopvQLHO https://t.co/h9vhVlsYnW
— Nick St. Pierre (@nickfloats) Apr 5, 2023
Headed home to NYC next week 🗽 Instead of doing a "who should I meet" tweet, here's my Top 10 food / drink spots I've accumulated over the last 25 years in NYC. 10/ Angel's Share / Pegu Club. Some of the best cocktails in town and immaculate vibes. RIP Pegu https://t.co/O5TO1WQSa1
— Robin Guo (@zebird0) Apr 5, 2023
This is a leap in image/pixel segmentation. Meta AI just released SAM (Segment Anything Model). One of the most interesting things is well understating of objects ("objectification" of parts). The model is released open source under an Apache 2.0 license, and it's only 2.4Gb.… https://t.co/LXdOWNqNTa https://t.co/SfX9KlcTbu
— Marco Mascorro (@Mascobot) Apr 5, 2023
Midjourney, the image AI, released a description system - it makes prompts for pictures so they can be reproduced. It shows how prompt-engineering alone will never be enough. Who would have described the pictures in the correct way (Neue Sachlichkeit?) AI will help us prompt AI. https://t.co/zkbYSa5Nmy
— Ethan Mollick (@emollick) Apr 4, 2023
Jarvis is now reality. My GPT-4 coding assistant can now build & deploy brand new web apps! It initializes my project, builds my app, creates a GitHub repo, and deploys it to Vercel. All from simply using my voice. https://t.co/0D5UbOjT5s
— Mckay Wrigley (@mckaywrigley) Apr 3, 2023
I love origami. I tried to create a perfect origami prompt for you. I hope you'll share your creations with me, sometimes it takes a few re-rolls but the results are so fun. Prompt: origami adorable orange rabbit, skyrim, kawaii, paper craft, cinematic, white background, high… https://t.co/Bam9LKx8mI https://t.co/rjbyu9HLYv
— Kris Kashtanova (@icreatelife) Apr 3, 2023
A Survey of LLMs A new 50 pages survey on large language models just dropped on arXiv. https://t.co/Tq2hBKUtNt https://t.co/86wEDNRlIV
— elvis (@omarsar0) Apr 3, 2023
MIT 6.S191: Recurrent Neural Networks, Transformers, and Attention https://t.co/rZAaElxH9F (https://t.co/EuoGiwnGZ3)
— Hacker News 200 (@newsyc200) Apr 2, 2023
With the sudden value of proprietary training data for AI… I am starting to think that all of us academics (and maybe the world overall) are going to regret handing over copyright to Elsevier & other publishers for every major academic paper written in the last half century. https://t.co/eFzxvM9sEr https://t.co/5kQz1gvbqE
— Ethan Mollick (@emollick) Apr 1, 2023
The most interesting AI news of the week for me is Bloomberg’s 50B parameter model trained on financial data. Points to a polyglot future where a number of players can win in AI, as opposed to just Big Tech and OpenAI.
— Matt Turck (@mattturck) Apr 1, 2023
Over past 20 yrs the #China Belt & Road Initiative #BRI has deployed >$240B in “bailout lending” to developing countries — with >$100B in just the past 3 yrs. This is primarily an effort to help prop up Chinese banks that made #BRI loans, which put their balance sheets at risk. https://t.co/zs34FQyVPM https://t.co/XYdxdM6fDx
— Dave McClure (@davemcclure) Apr 1, 2023
LLaMA 30B using only 5.8GB of ram. Will be very interesting to see some benchmarks on this, feels too magical https://t.co/DbvuvDWfc8
— anton (@abacaj) Mar 31, 2023
