- Lumian
- Posts
- Lumian Gen AI Newsletter Issue #36
Lumian Gen AI Newsletter Issue #36
Meta’s AI Clusters, BharatGPTs Hanooman, Midjourney’s Consistency
Robin Lobo & Saswat Panda
March 15, 2024
Welcome to the 36th edition of the Lumian Weekly Gen AI Newsletter!
Alright, let's get into it. A young start-up, Cognition Labs, has just rolled out Devin — an AI Software Engineer — and let me tell you, it's not your average code-monkey. Imagine an AI that doesn't just correct your semicolons but takes your vague ideas and turns them into fully-functional software while you're off doing... well, whatever it is you do when you're not coding. And it does so with a flair that makes conventional coding assistants look like they're still figuring out "Hello, World!"
So, what's the big deal? Behind this wizardry is a team so decked out with International Olympiad in Informatics gold that they could open their own jewelry store. And with $21 million in Series A funding, they're playing to win.
Now, I hear you saying, "But jobs! The robots are coming for our jobs!" Well, yes and no. Sure, some roles will transform or fade into the sunset, but software engineering isn't going away; it's just getting a futuristic facelift. We're looking at a world where AI doesn't replace us but makes us look like rockstars (or, at the very least, highly efficient coders).
So, whether you're a developer, a dreamer, or just someone who appreciates a good tech revolution, Devin is a sign of the times. Embrace it, explore it, and who knows? Maybe you'll find yourself working alongside an AI colleague sooner than you think. Welcome to the future—it's looking pretty clever.
As always, it’s been a busy week and we’ve got it all covered for you!
Happy reading! 📚🤖🎵
In this week’s issue:
News Flash: Meta’s AI Clusters, BharatGPTs Hanooman, Midjourney’s Consistency
AI Frontier: AI Market Research tools you can use today
Fundraising: The biggest deals in AI
Nerd Out: Technical and Business Content for Everyone
⏱️ News Flash
The 2-Minute Scoop to Keep You in the Loop
Meta introduced two large AI training clusters with over 24,000 GPUs each, aiming for 600,000 GPUs by end of 2024, to power advanced AI models like LLaMA 3. They're using open-source hardware and software, experimenting with networking technologies to improve performance. This move not only boosts their AI research capabilities but also aims to attract top talent by being transparent about their projects, including the development of LLaMA 3.
The BharatGPT group, backed by heavyweight partners, has launched Hanooman, a ChatGPT-like series capable of understanding and interacting in 11 Indian languages, promising to revolutionize sectors like healthcare, governance, financial services, and education. Despite challenges in dataset quality, notably the reliance on synthetic datasets, the initiative marks a significant step towards inclusivity and technological advancement in AI for India. The first four open-source models in the series (1.5 billion, 7 billion, 13 billion, and 40 billion parameters) are set to be released next month.
Midjourney has introduced a new feature that allows for the consistent recreation of characters across different images by using a character reference tag (“–cref”) along with a URL to a previously generated image of the character. This addresses a significant challenge in AI-generated imagery by enabling characters to maintain consistent appearance across various scenes, crucial for storyboarding and narrative continuity.
🚀 AI in Practice
Cutting-Edge AI Market Research Tools You Can Use Today
🤑 Fundraising
The (AI) Intelligent Investor
🤖 Nerd Out
Technical and Business Readings
😜 AI Art!
Inspiration or Plagiarism?!
How did you like this week's newsletter?Vote below: |
If you were forwarded this newsletter, you can access more of our content by subscribing here.
Best,
Reply