Lumian Gen AI Newsletter Issue #34

⏱️ News Flash

The 2-Minute Scoop to Keep You in the Loop

1/ Scale AI signs a 1-year contract with the Pentagon

The Pentagon has teamed up with Scale AI to evaluate how well AI technologies can handle military tasks, focusing on understanding military language and scenarios. By developing a rigorous testing process, they aim to identify the strengths and limitations of AI in defense applications.

2/ OpenAI’s fundraising deal values it at $80B

OpenAI has finalized an agreement that values the firm at over $80 billion. This deal allows employees to sell their shares through a tender offer led by Thrive Capital, rather than raising new funds for the company. This move comes after a series of significant investments in OpenAI, including a notable $10 billion from Microsoft in early 2023.

3/ Reddit bets on AI ahead of its IPO

Reddit is preparing for its IPO by licensing its user-generated content to an unnamed AI company for model training, securing a $60 million annual deal. This strategic move follows Reddit's previous tightening of data access, aiming to monetize its vast data resources more effectively. The agreement signals a trend where social media platforms control and profit from their data, potentially shaping the future interaction of AI chatbots and user engagement with technology.

🤖 Diversity in AI?

Google Hits Pause on AI Image Generation

In the latest episode of "AI Does the Darndest Things," Google found itself apologizing for the antics of Gemini, which, in a bold move, decided to remix history by generating images of the U.S. Founding Fathers and WWII-era German soldiers as a diverse cast of characters. It's as if Gemini took one look at the history books and thought, "Needs more diversity. I can fix that."

Google issued an "Our bad" and hit the pause button on Gemini's people-generating prowess. They promised to come back with an "improved" version, presumably one that knows not to give certain parts of history a diversity makeover.

This comes at a time when Google's AI, in a race against Microsoft-backed OpenAI and its shiny new Sora, is trying to prove it's the smart kid in class by not just spitting out words but also pictures. Last week, Sora was showing off, turning text prompts into videos, while Gemini was left explaining why its depiction of history looked like a casting call for a very progressive time-travel movie.

The kerfuffle has reignited discussions about AI and bias. It's a reminder that when it comes to AI, maybe we're still figuring out the difference between making history and remaking it.

🤑 Fundraising

The (AI) Intelligent Investor

• BRIA raises $24M in a Series A for responsible AI images

• Bioptimus raises $35M seed round to develop AI foundational model focused on biology

• Self-driving startup Recogni raises $102M to design AI accelerator chips now

🤖 Nerd Out

Technical and Business Readings

• Inside the funding frenzy at Anthropic

• You should be playing with GPTs at work

• Why the New York Times might win its copyright lawsuit against OpenAI

😜 The AI Peace Treaty!

Technological Irony!

If you were forwarded this newsletter, you can access more of our content by subscribing here.

Best,

Robin and Saswat