S¹⁹⁸ubscribe¹⁹⁸ | Ben’s Bites Pro¹⁹⁸ | Ben’s Bites News⁶⁸
Daily Digest #432

Want to get in front of 100k AI enthusiasts? Work with us here^89†

Hello folks, AGI needs two essential elements: core research and funds to bring to everyone. That’s what we have today;

PICKS

1. Mistral AI has raised €600M in Series B funding.¹⁷⁵ General Catalyst led this funding (their thesis²²⁰)—valuing the company at $6B. Quick reminder: Mistral announced their seed funding (whopping $113M) just a year ago and they’re one of the top LLM makers in the market now. We’ve done a 🍿case study on Mistral’s journey²⁰⁰ before.

2. “LLMs are not the way to AGI.”⁴²³ There’s a large camp of AI researchers who believe this to be true. But what to do about it? François Chollet, the creator of Keras and Mike Knoop, co-founder of Zapier are launching a $1M challenge to find alternative ways to get to “general intelligence”. Hear ’em talk on Dwarkesh’s Podcast²⁰⁶ and the No Priors Pod¹³³ or… you can read 🍿Our Summary¹² (also below).

3. Microsoft is retiring the GPT Builder³³⁶ for its consumer customers. Shocker!! GPTs are useful, but Microsoft and OpenAI's commitment to them as a product has been shaky. Microsoft’s GPT Builder will still be available for enterprise customers and this current news doesn’t concern ChatGPT’s GPT store in any way.

from our sponsor

The old way of delivering customer service is dead. That’s a good thing. 

The new way is AI-first. 

AI-first customer service is better, faster, and cheaper.

With Intercom’s complete AI-first customer platform⁶⁶:

• Customers get instant, accurate answers

• Agents get more time to solve the complex problems

• Leaders get a cost-effective solution

Learn more about AI-first customer service without compromise.⁸⁷

TOP TOOLS

• Intercom⁹³ - Boost agent efficiency 31% with Fin AI Copilot¹⁷⁴, a personal AI assistant*

• Model Personalization in Midjourney²³⁹ - Tune the MJ algorithm to your own personal tastes (demo²¹¹).

• Teamcreate⁴²³ - Build your dream team in minutes.

• Galileo AI⁶⁰⁹ - Go from idea → design → code in under a minute.

• Flyhomes AI ⁴⁰²- World’s first AI-powered home search (and research) portal.

• Deformity³⁸¹ - Conversational forms that wow respondents.

• Soloist⁴¹⁸ - Create your website from a few simple inputs.

• Namify AI²⁹³ - Free AI-powered business name generator.

View more →¹⁴⁴
*sponsored

NEWS

• Elon Musk drops suit²⁵³ against OpenAI and Sam Altman.

• Mira Murati fires back⁴⁵⁸ as Elon calls the OpenAI-Apple partnership a “creepy spyware”.

• OpenAI partners with Oracle¹¹⁴ to extend its inference infrastructure.

• Brazil hires OpenAI²²⁶ to cut costs of court battles.

• Google announces LLMs and agents focused on personal health.³⁵⁸

• AI is getting very popular among students and teachers³⁴⁰, very quickly.

• American AI startups attempt to relocate China-based engineers.¹²⁸

• Particle raises $11M Series A¹⁵⁰ to accelerate development of AI-powered news app.

• Linq raises $6.6M¹³³ to use AI to make research easier for financial analysts.

• Cogniny raises $100M²⁰² Series C for its AI agent workforce.

View more →⁸⁷

Unclassifieds - short, sponsored links

• Intercom³⁵ - Deliver better, faster, cheaper customer service with the only complete AI-first platform. Learn more.⁸⁷

QUICK BITES

“LLMs are not the way to AGI.” There’s a large camp of AI researchers who believe this to be true. But what to do about it? François Chollet, the creator of Keras and Mike Knopp, co-founder of Zapier are launching a challenge⁴²³ to find alternative ways to get to “general intelligence”.

What is going on here?

ARC Prize is offering $1M to improve the reasoning of AI models (not just LLMs).

What does this mean?

ARC stands for Abstraction and Reasoning Corpus. It’s a collection of tasks where you have to observe a few input-output pairs and predict the output for a similar input. We humans are pretty good at it.

Just try this example. You can easily guess the output for the Test Input.

SPOILER: Just get rid of the smaller piece. Play it here¹⁹⁹ and guess the colour.

But what if AI systems try it? Well, the current state-of-the-art performance is 34% accuracy (humans are at 85%). So, the ARC-AGI challenge thinks this is a good challenge to test general reasoning.

It defines AGI as a system that can efficiently acquire new skills outside of its training data. And to top it all, there’s a million-dollar prize pool for beating humans at this challenge.

François and Mike talk about the challenge more on Dwarkesh Patel’s podcast²⁰⁶ and the No Priors Pod.¹³³ 

Why should I care?

The explanation by the organizers makes sense. LLMs don’t reason, instead, they copy-paste reasoning patterns in the vast amount of data they consume while training.

The organizers also argue that the progress in AI has turned closed source and the narrative of “scale is all you need” is hurting the direction of global research. New ideas are needed and I’m up for it (except that I spent way too much time playing the puzzles last night).

Share this story¹²

Ben’s Bites Insights

We have 2 databases that are updated daily which you can access by sharing Ben’s Bites using the link below;

• All 10k+ links we’ve covered, easily filterable (1 referral)

• 6k+ AI company funding rounds from Jan 2022, including investors, amounts, stage etc (3 referrals)