Lumian Gen AI Newsletter Issue #8

⏱️ News Flash

The 2-Minute Scoop to Keep You in the Loop

1/ Meta launches a new speech-generating AI model

Meta revealed an AI marvel, "Voicebox", a text-to-speech tool, crafting voices from scratch and modifying samples. It is multilingual in six languages and 20 times faster than its predecessors. It doesn't even need a specific voice to train on. Imagine snipping a dog's bark from a video and seamlessly replacing it with crystal-clear conversation! But you can see why Meta might be reluctant to let this genie out of the bottle.

2/ E.U. Parliament approves AI Act that challenges tech giants

The EU's charging forward with a new AI law, seemingly in a race to out-regulate the US, China, and Japan. While we’re all for boundaries, we can't help but worry that this might hit our EU AI firms, maybe even slow down innovation. We're talking potential fines for open source misuse and tighter controls on APIs. We hope that these rules focus more on robust systems and human adaptation, and not be just another “accept cookies” banner.

3/ The Awkward Partnership Leading the AI Boom

Looks like the AI dream team, Microsoft and OpenAI, are hitting some bumps in the road. Microsoft's unconventional deal (snapping up a 49% stake in OpenAI's profit-making wing) has been more of a head-scratcher than a high-five, sparking clashes and confusion. From murky model architectures to customers juggling sales pitches from both sides, and Microsoft splitting its dollars between in-house AI and OpenAI - it's a messy affair. This offbeat alliance could redefine the startup-big player relationship for years to come.

🖼️ McKinsey’s Gen AI Report

The Economic Potential of Generative AI

McKinsey's new report is a goldmine on the economic power of generative AI. We’ve summarized it to its essence but highly recommend the full read.

The report predicts that Gen AI could crank up the total AI impact by 15-40%, injecting an annual $2.6T to $4.4T across 63 applications. It identifies four hotspots capturing 75% of the value:

• Customer operations

• Marketing and sales

• Software engineering

• R&D

It highlights that Gen AI could flip work on its head, automating 60-70% of tasks and predicts a hike in labor productivity by 0.1-0.6% yearly through 2040, based on tech uptake and how work time is reassigned.

Of course, there'll be hurdles including:

• Managing the risks inherent in generative AI

• Determining what new skills and capabilities the workforce will need

• Rethinking core business processes such as retraining and developing new skills

If played right, Gen AI could fuel economic growth and a fairer, greener world. Buckle up for a wild ride!

🤑 Fundraising

The (AI) Intelligent Investor

• Accenture looks to power AI efforts with $3 billion investment

• Mistral AI raises $113M Seed round at a $260M valuation to take on OpenAI

• Synthesia raises $90M in Series C for professional AI videos

🤖 Nerd Out

Technical and Business Readings

• Demystifying GPT Self-Repair for Code Generation

• GPT-4 outperforms humans in pitch deck effectiveness

• Market Map: Gen AI Companies with Foundational Models

😜 Fundraising for AI Start-ups!

Time to build the moat!

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

Best,

Robin and Saswat8