- Lore
- Posts
- Lore #80: Weather Forecasting, Gene Editing, and Speech AI
Lore #80: Weather Forecasting, Gene Editing, and Speech AI
Good morning,
Another week with huge acceleration.
Here are the highlights:
Weather Forecast Revolution: Google's SEEDS AI model promises to predict severe weather faster and more economically.
Gene Editing Advances: Profluent has developed a way to edit DNA with AI using CRISPR technology, potentially greatly enhancing genetic engineering capabilities.
Stable Diffusion 3 Released: Stability AI finally released its latest open-source Stable Diffusion models.
Also, here is the latest episode of The Next Wave, discussing the positive potential and dangers of Speech AI that will be made possible by technology such as Hume AI.
NEWS
Google has developed a new AI model, the Scalable Ensemble Envelope Diffusion Sampler (SEEDS), which promises to deliver faster, more economical weather forecasts. This generative AI model is inspired by the architecture of large language models and can quickly generate multiple weather scenario ensembles. SEEDS is designed to enhance the prediction of extreme weather events, which are typically challenging and costly to forecast with traditional physics-based models. By efficiently producing up to 31 predictive ensembles from minimal input data, SEEDS aims to improve the accuracy and speed of weather forecasting, potentially saving lives by providing earlier warnings for severe weather conditions.
Generative AI has entered the realm of gene editing, promising to revolutionize the field much like it has with digital content creation. A new AI system developed by Profluent is creating blueprints for mechanisms capable of editing DNA.
These developments, which Profluent will present at an upcoming American Society of Gene and Cell Therapy meeting, leverage the same type of technology that powers models like ChatGPT, but are applied to generating novel CRISPR-based gene editors after extensive analysis of biological data. Profluent has taken a bold step by open sourcing one of these tools, OpenCRISPR-1, allowing widespread experimentation and potentially accelerating innovation in gene editing.
Stability AI has launched Stable Diffusion 3 and its enhanced version, Stable Diffusion 3 Turbo, on its Developer Platform API in partnership with Fireworks AI. This new release aims to match or surpass other leading text-to-image models like DALL-E 3 and Midjourney v6 in terms of typography and prompt adherence. Featuring a new Multimodal Diffusion Transformer architecture, the models offer improved text understanding and spelling capabilities.
Other news
OpenAI’s Sam Altman is funding a green-energy moonshot as AI’s power demands grow to ‘insatiable’ levels (Link)
Microsoft introduces smaller AI model (Link)
Meta, in its biggest AI push, places smart assistants across its apps (Link)
How robots are taking over warehouse work (Link)
Olympic organizers unveil strategy for using artificial intelligence in sports (Link)
Google 'considers charging' for premium search results powered by artificial intelligence (Link)
Unleash the Power of AI with Graydient
Step into the future of AI with Graydient, your ultimate destination for unlimited Stable Diffusion and unfiltered language models on all your devices.
What Graydient AI offers:
Unlimited Credits: Enjoy endless AI creations without worrying about running out of credits.
Preloaded Models: Access 8 terabytes of the best AI models, ready to use without any boot time.
Multi-Device Compatibility: Experience lightning-fast performance on your PC, Mac, tablet, or laptop.
Diverse Applications: Explore a wide range of AI-powered tools for image generation, chatbots, and more.
Community Engagement: Connect with like-minded AI enthusiasts in our vibrant interest groups.
Flexible Plans: Choose from Starter, Pro, and Plus plans to suit your needs and budget.
As an exclusive offer, sign up now and get 2 months free, plus save 15% off your subscription. Don't miss out on this opportunity to join the AI revolution with Graydient.
Join Graydient today and embark on a journey of limitless AI possibilities. Plus by signing up, you help support this newsletter which we work hard on. Thank you!
INSPIRING PROGRESS
Teaching machines how to see
at TED 2024
"540M years ago, there were no eyes. Light went unseen. Trilobites were the first to sense light, and the Cambrian Explosion of body plans followed. Sight led to insight. Seeing led to understanding and action, the elements of intelligence. We are… twitter.com/i/web/status/1…— Steve Jurvetson (@FutureJurvetson)
10:45 PM • Apr 21, 2024
Microsoft’s VASA-1 generates realistic video from a single photo
The First AI-Generated Video That Looks Super Real
Microsoft Research announced VASA-1.
It takes a single portrait photo and speech audio and produces a hyper-realistic talking face video with precise lip-audio sync, lifelike facial behavior, and naturalistic head movements… twitter.com/i/web/status/1…
— Bindu Reddy (@bindureddy)
11:17 PM • Apr 17, 2024
GPT-4 makes huge leaps in medicine, even outranking physicians in some cases
“GPT-4 ranked higher than the majority of physicians in psychiatry… it performed similarly to the median physician in general surgery & internal medicine… GPT-4 performance was lower in pediatrics & OB/GYN but remained higher than a considerable fraction” of active doctors.
— Ethan Mollick (@emollick)
11:03 PM • Apr 16, 2024
Meta's Segment Anything Model (SAM)
Meta's Segment Anything Model (SAM) can now run in your browser w/ WebGPU (+ fp16), meaning up to 8x faster image encoding (10s → 1.25s)! 🤯⚡️
Video is not sped up! Everything runs 100% locally thanks to 🤗 Transformers.js and onnxruntime-web!
🔗 Demo: hf.co/spaces/Xenova/…
— Xenova (@xenovacom)
12:44 AM • Apr 21, 2024
Boston Dynamics’ new and improved all-electric Atlas robot
Boston Dynamics reveals its new, improved all-electric Atlas robot, and @bheater has all the details right here: tcrn.ch/3vKUeeN
— TechCrunch (@TechCrunch)
1:17 PM • Apr 17, 2024
FURTHER STUDY
My mind is blown.
@GroqInc is serving LLaMA 3 at over 800 tokens per second!
800. Tokens. Per. Second.
This unlocks so many incredible use-cases.
It's one thing to see my demo — it's another thing entirely to experience it for yourself.
Do yourself a favor and try it asap. twitter.com/i/web/status/1…
— Matt Shumer (@mattshumer_)
4:13 PM • Apr 19, 2024
VENTURE DEALS
Rivos raised $250M to create an accelerator for generative AI and data analytics workloads (Link)
Greeneye Technology raised $20M for precision spraying of crops (Link)
Mentee Robotics raised $17M for its humanoid robot, MenteeBot (Link)
Vidyard raised $15M for personalized AI avatars (Link)
AI Squared raised $13.8M to deliver data and AI insights into business applications (Link)
FYLD raised £12M for its field management platform (Link)
Veesual raised $7.5M for virtual clothing try-on tech (Link)
AIPERIA raised €7.5M for food waste reduction (Link)
Somite raised $5.3M to incorporate AI in stem cell therapy (Link)
HealthSage AI raised €3M for healthcare delivery and patient pathways (Link)
Spatialedge raised $3.1M to help companies build machine learning solutions (Link)
MokSa AI raised $1.5M to curb shoplifting with AI cameras (Link)
Coming up:
France's Mistral AI seeks funding at $5B valuation (Link)
That's it!
See you next week!
If you want to support Lore, please:
Share this newsletter
Follow us on Twitter @NathanLands & Lore.