• Ben's Bites
  • Posts
  • Daily Digest: New CodeLlama is better than GPT-4

Daily Digest: New CodeLlama is better than GPT-4

PLUS: family friendly AI and AWS hackathon.

January 30, 2024

Hello folks, if you work at Meta, Microsoft, Google, Apple or Amazon - I’d love to know if your company is doing anything internally to encourage AI adoption for employees; like hackathons, workshops, internal tools etc - reply here 🙂 I’m writing a post about it.

here’s what we have today;

PICKS

  1. Meta releases Code Llama 70B, a larger (and better) version of their code generation models. It achieves 67.8 on HumanEval benchmark which is higher than GPT-4’s base rating of 67. And again it’s open source with commercial use allowed. You can use the bigger models now on Huggingface, Perplexity or locally with Ollama.

  2. Common Sense Media and OpenAI are shaking hands to shape the future of AI for youngsters and their families. There are two goals of the partnerships: teach families to get the most out of it and deal with any risks by exposing kids to it.🍿Our Summary (also below)

  3. AWS is organising a hackathon on its AI app-building playground called PartyRock. The deadline is 12th March and the top prize is $20,000 in AWS credits (with multiple other prizes in the few thousand $$ range, both cash and credits).

TOP TOOLS

  • AI Autocomplete by Shortwave - Finish your sentences based on your previous emails.

  • Supadash - AI-generated dashboard and charts from your database.

  • Daydream - BI for C-level, finance and ops, with AI.

  • Whisperfusion - Seamless conversations with AI with ultra-low latency.

  • Podnotes - Turn podcasts into content with AI.

  • Meetrics - Transcription doesn't solve bad meetings, meeting prep does.

  • Unfetch - Create AI workflows with any API.

  • Screenshot Save - Rename all your desktop screenshots with AI.

NEWS
QUICK BITES

Common Sense Media and OpenAI are shaking hands to shape the future of AI for youngsters and their families. There are two goals of the partnerships: teach families to get the most out of it and deal with any risks by exposing kids to it.

What is going on here?

OpenAI and Common Sense Media partner to craft AI that's friendly for families and teens.

What does this mean?

Some context: Common Sense Media is a nonprofit creating advisory content for parents and kids to navigate the digital media landscape.

OpenAI and Common Sense Media are rolling out guidelines and educational stuff for the folks at home and in schools, ensuring everyone's up to speed on using AI without any hiccups. Plus, they're curating a section in the GPT Store that's all about family-friendly content, vetted by Common Sense's own standards.

Talking about this at the Common Sense Media summit, Sam Altman (CEO, of OpenAI) thinks that the idea that AI is bad for kids is flawed. There are risks, and OpenAI is anxious about them but to not teach them these tools is a bigger mistake. And about kids getting lazy? Altman has said in past: We adapted to calculators and changed what we tested for in math class.

Why should I care?

Common Sense Media has an AI section already with guidelines and ratings for top AI providers. But this partnership goes beyond putting up a stamp: “Hey, this is risky, don’t use it”. Both OpenAI and Common Sense Media are trying to create resources that families, schools and teens can actually use. And I guess that’s a step in the right direction.

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)

Reply

or to participate.