Children Need Help with GenAI, Pizza Video Challenge Update, and the COPIED Act

A few links to kick off this artificial intelligence week.

Jul 15, 2024

Morning y’all!

If you’ve been unplugged from the internet this weekend then you may have missed the insanity that is a possible assassination attempt on former-President Trump.

This will never be anything remotely close to a political newsletter but something radical could have changed US Politics and that would have also impacted the broader technology industry (and thus us as well).

Wow. Wow. Wow. I have to try to focus today!

※\(^o^)/※

— Summer

Here’s a list of 28 AI startups that have raised more than $100m in 2024.

Whoa. This is good news for those that care about safety:

A research team at the University of Oxford has introduced a new method called "Semantic Entropy Probes" (SEPs) to efficiently detect uncertainty and hallucinations in large language models. The method could make the practical application of AI systems safer.

Note that I said “those that care” since there are many folks who really don’t care about it at all and, to be honest, I’m not sure I care as much as I used to. Don’t hate me here folks! It’s just what it is.

Min Choi used Claude 3.5 Sonnet to create Gen-3 Alpha prompt builder. What.

Tomasz has some thought on AI Agents and software pricing, including the idea of triple the pricing. Well, that’s one way of doing it.

So where are we with the pizza AI challenge? Prototypes are faster but quality is still sus and will take more time.

I’m not sure how to feel about this:

A study by researchers from the Universities of Washington and Michigan explored how children aged 7 to 13 use generative AI tools like ChatGPT and DALL-E for creative tasks.

Do children really need help navigating generative AI?

Whistleblowers just filed a complaint with the SEC alleging that OpenAI used overly restrictive non-disclosure agreements to prevent employees from reporting concerns to regulators, violating federal whistleblower protections.

Now this is neat:

U.S. Rep. Jennifer Wexton debuted an AI-generated version of her voice, allowing her to continue addressing Congress despite speech limitations caused by a rare neurological condition.

Open source issue tracking with AI.

Instantly provide automated support to customers and internal teams with AI agents trained on internal knowledge. Set up effortlessly on popular collaboration platforms.

I have a friend in this world who’s built his first VC firm recently and so I send him apps like this to review; they just raised a round too.

The COPIED Act seems to protect journalists and artists from having their work used by AI models without consent. Removing watermarks would also become illegal.