Meta's New Image AI

PLUS: EU passes AI Act


Good morning, human brains. Welcome back to your daily munch of AI news.

Here’s what’s on the menu today:

  • AI Cupid? 💘 

    Is AI the Tinder hack we’ve all been looking for?

  • Meta releases human-like image AI 🖼️

    Legend Yann LeCun is the architect behind the new model.

  • The EU passes the EU AI Act 🧑‍⚖️ ⚖️ 

    The law has spoken!


Cupid from the machine? ❤️

Lucy Cavendish from the Daily Mail let artificial intelligence take over her love life.

Yep, you read that right.

This brave soul handed over control to Genie, the AI advice app. The AI suggested message approaches to Lucy’s Tinder matches based on profile bios.

It also gave its opinion on the lighting of profile pictures, what to wear on the first date, and even on the tone in her messages.

What Lucy wanted to wear, compared to Genie’s outfit.

And it didn’t stop at the initial, flirty Tinder messages. Genie gave real-time advice during her real-life dates. From “is ordering a 3rd glass of wine is rude?” to “what can I say to politely leave the date?” the bot had Lucy’s questions covered.

The downside? She spent a lot of time in the loo using her phone.


So, did it work?

First dates were plentiful, but there were no follow-up dates after that. Between the AI’s boundary-setting behavior and lying about shared interests with her potential partners, it was difficult to make an authentic connection.

Our take: Dating is plenty full of mind games and nuances. Is it really necessary to add more confusion with AI? Is your date legitimately interested in your passion for aquatic invertebrates, or is it just AI leading you along?


Reinforcement Learning

Reinforcement learning refers to a type of machine learning where the model under training receives feedback through positive or negative reinforcement to improve over time.

The process more closely resembles trial and error as opposed to other methods, like supervised or unsupervised learning.


An Inbox Built For Newsletters.

Bot Eat Brain looking darn good in the Apiary inbox.

Newsletters rock, but email sucks.

Your inbox wasn’t built for newsletters. It was built for a note from your Mom, an update from your boss, and a notification that your package was delivered.

Back in the day, you knew every email you got mattered.

Now? The few emails you want to read are buried under mounds of sales emails, coupon codes, and random junk.

You deserve better. Take back your inbox with Apiary, the email reader built for newsletters.

No more clutter: Add your favorite newsletters to Apiary and see only those newsletters.

No more missed emails: Apiary filters out the noise and lets you know when the good stuff arrives.

Learn 10x faster: Apiary lets you take notes directly in the newsletter and never forget key insights.


Meta releases ‘human-like’ image AI 🖼️

Their new model, I-JEPA, analyzes and fills in missing pieces of images.

What’s so great about it?

It uses its worldly background knowledge to fill in the gaps instead of just relying on neighboring pixels.

With the new approach, we can bid farewell to all those extra AI-generated fingers… hopefully.

The model is based on a new architecture proposed by Meta’s Chief AI Scientist, Yann LeCun. A legend in the field, he’s considered to be the “godfather of AI”.

Drawing controversy, Meta execs refused to sign a statement likening AI risks to pandemics and wars — backed by other large players like OpenAI, DeepMind, Microsoft, and Google. LeCun himself has been outspoken against “AI doomerism”, proposing built-in safety checks as the right path forward.

You can read the full paper behind the model here.

Our take: Cool new image tech? Check. AI safety drama? Also check. Given this first success, it’s likely we’ll see more models using this new architecture.


The European Parliament passes the EU AI Act 🧑‍⚖️ ⚖️ 

The European Parliament just gave a big thumbs up to the EU AI Act.

499 MEPs voted in favor, with 28 disagreeing and 93 abstaining.

The landmark decision has been months in the making, with plenty of involvement from players like OpenAI and Google.

And it certainly looks like OpenAI and others are affected by the new transparency regulations.

Our take: AI development in Europe is about to look a lot different. Big AI is on move — what’s next?



Startup News

Mistral AI secures a whopping $113 million in seed funding with notable investors including Lightspeed, Redpoint, Index Venturesa and Xavier Niel.

UK universities receive £54 million from gov — with collaboration across academia, business, and the wider public, the initiative aims to foster responsible and trustworthy AI research.


Alzheimer's AI research on brain fluid flow — a major breakthrough with far-reaching implications for treating diseases like Alzheimer’s.


Lovo — Generate lifelike human voices for podcasts or videos. [Sponsor] — Websites with SEO-ready and speed-optimized content built in 60 seconds.


There are a few AI tools out there… to say the least:


If you like Bot Eat Brain, you might like these other newsletters too:

🏥 Healthcare AI News - 5 minutes or less is all it takes to elevate your Healthcare knowledge with this expert-curated weekly AI newsletter. Stay informed and stay ahead.

🖊️ The Pen Pivot — AI tools and prompts curated for curators.

☀️ The Boonly — Brighten your inbox with inspirational insights that make self-growth enjoyable, not stressful. Every Sunday.


Until next time 🤖😋🧠

What'd you think of today's newsletter?

Login or Subscribe to participate in polls.