Double win for Open AI

PLUS: Stability AI scores as well, Amazon looks back at Alexa

May 01, 2023

Hey folks, welcome to 5731 of you who joined last week. We are officially past 90,000 subs. Insane, and thanks for reading.

Today we have new models from Stability AI, Amazon’s plan to make Alexa great again, and Open AI’s funding plus Italy conquest.

Let’s go.

From our sponsor:

Apply to Tofth School, a ten-month program designed to bring into view the creative powers of contemporary technology – from AI to Synthetic Biology. 

The program is free to attend, comes with stipend-based support of up to $50,000, and includes a world-class faculty of leading industry names. Successful candidates have the chance to pursue building products and companies with the support of ToftH's network.

🤌 Our Picks

Highlights if you've only got 2 minutes

1/

Stability AI are really hitting the ground running at the moment - we’ve got a double pick for them today with two model releases. First up is a new text-to-image model they’ve named DeepFloyd IF, and it really is the moment we’ve all been waiting for - it can actually do a half decent job of rendering text in generated images! Finally! (No news on fingers though…) Secondly, they’ve open-sourced a language model called StableVicuna, claiming that it’s the first large-scale instruction fine-tuned and RLHF model. Available on Hugging Face too!

2/

Well if this isn’t overdue then I don’t know what is - Amazon’s Alexa might be about to get the upgrade we’ve all been expecting since language models really hit the scene last year. CEO Andy Jassy is rallying behind the idea of building the world’s best personal assistant, and Alexa is the obvious place to start. It can be such a frustrating thing to use at times, and using the current powerful state-of-the-art language models could be just the kick it needs.

3/

As if that $10 bill from Microsoft wasn’t enough, OpenAI have just gone and got $300m from a bunch of VCs - that gives them a valuation of close to $30 billion… On another OpenAI note, they’ve managed to win the Italians over! After getting a to-do list from the tech watchdog there, OpenAI duly complied and ChatGPT is back up and running in the Med. They’ll have to provide training data info, will have an opt-out form for using your data for training, and have an age confirmation box. I wonder if their willingness and speed with complying will encourage more countries to follow in Italy’s footsteps?

4/

Quick one: AutoGPT has launched a waitlist for their GUI. (link)

🛠️ Cool Tools

Product launches, updates and demos

  • Lamini - Enabling engineering teams to outperform general purpose LLMs through RLHF and fine-tuning. (link)

  • MLC LLM - Run an LLM locally with no server support and accelerated with local GPUs on your phone and laptop. (link)

  • Play ht launches multilingual synthesis and cross-language voice cloning. (link)

  • AutoPR - Autonomously write pull requests in response to issues with ChatGPT. (link)

  • Guidde AI - Create video documentation instantly with the magic of AI. (link)

  • BabyBeeAGI - A buggier, slower, but more powerful mod of BabyAGI. (link)

  • Hackercast - An AI-generated podcast summary of Hacker News. (link)

  • Plugin to help ChatGPT to interact with IoT devices. (link)

  • Monterey v1.0 - Copilot for product insights. (link)

  • Trumpet House - Unleash your Twitter potential with AI. (link)

  • Emdash - Use on-device AI to learn more from your book/article highlights. (link)

  • Lettria - Text processing for everybody. (link)

  • LLM report - An OpenAI API analytics dashboard. (link)

🤓 Miscellaneous

News, podcasts, videos, blogs etc

  • Mapping the future of truly open models and training Dolly for $30. (link)

  • Navigating the high cost of AI compute. (link)

  • AI has hacked the operating system of human civilization - Yuval Noah Harari. (link)

  • New layer of abstraction, not the end of software engineering. (link)

  • In AI who wins? with Sarah Guo. (link)

  • Pilot program aims to see if AI will cut time that medical staff spend replying to online inquiries. (link)

  • Text-to-audio generation using instruction-tuned LLM and latent diffusion model. (link)

  • Five worlds of AI. (link)

  • Interviews with EU lawmakers who wrote a draft bill on generative AI in just 11 days. (link)

  • The future of generative AI is niche, not generalised. (link)

  • Elon Musk ramps up AI efforts, even as he warns of dangers. (link)

  • I decided to outsource my entire personal financial life to GPT-4. (link)

  • From scams to music, AI voice cloning is on the rise. (link)

  • How AI is already changing the 2024 election. (link)

  • ChatGPT outperforms physicians in quality, empathetic answers to patients. (link)

  • WSJ journo challenged AI clone to replace her for 24 hours. (video, article)

  • We aren't close to creating a rapidly self-improving AI. (link)

  • The co-founder of Skype invested in some of AI’s hottest startups—but he thinks he failed. (link)

🎓 Learn

How-to’s and resources

  • Use ChatGPT as a search engine. (link)

  • A brief history of LLaMA models. (link)

  • Timeline for AI/ML papers with LLM/transformer models. (link)

👋 Too many links?! I created a database for all links mentioned in these emails. Refer 1 friend using this link and I'll send over the link database.

📰 Unclassifieds

Short, sponsored links 

  • Bring your ideas to life - no code required. Take No Code MBA's project-based courses creating apps like Airbnb, Reddit, and ChatGPT.

  • Revolutionize your blogging with Hashnode AI! Unlock powerful features to enhance writing & SEO. Try now 🚀

Have a product, service, job, event, newsletter, app, book, movie, tool, or anything you'd like to share with over 80k subscribers? 
Advertise

🖼 AI Images of the day

Funny memes and pics from around the web

Send this to 1 AI-curious friend and receive my AI project tracker database! Use this link.

How was today's email?

Login or Subscribe to participate in polls.

Reply

or to participate.