White House mega deal with AI companies

Hello folks, here’s what we have today;

 Our picks 

1/

The White House has seven AI companies to agree to abide by a set of asks on AI safety. These are Amazon, Google, Meta, Microsoft from the big tech, and Open AI, Inflection, and Anthropic from the AI front. These rules are only voluntary, so there won’t be any actual ramifications of ignoring them.

With the media attention and statements, this feels like they have really started getting the ball rolling on the AI safety and regulation front. Here are some of the official statements:

Google: Advancing bold and responsible AI
Inflection: The precautionary principle
Open AI: Moving AI governance forward
White House: Voluntary commitments from leading AI companies

2/

OpenAI is bringing the ChatGPT mobile app to Android users. It’s currently on the Play Store, where you can pre-register for auto-download when it opens up next week. Personally, the iOS app was a game changer for me. Hope it’ll be the same for Android users.

3/

Stability AI is keeping things coming, and this time we’re back on the language model front. FreeWilly 1 and 2 are their open access LLMs that they claim to have “exceptional” performance. They both rely on Meta’s LLaMA foundation models with a bit of supervised fine-tuning thrown into the mix, and even get to a point of matching GPT-3.5.

4/

The Open Source Initiative says that Meta’s LLaMa 2 license is not Open Source. They do praise Meta for making things easy to access but are salty about Meta calling it open source.

Here’s another POV about how LLaMA2 isn't open source—and why it doesn't matter. Aleesio’s breakdown of what different forms open source can take for AI is interesting.

from our sponsor

Generative AI alongside your DAM

Turbocharge your content creation with Bynder's new generative AI capabilities.

Our robust controls allow you to use AI on your own terms.

Here's how:

✅ Decide which content items can use AI assistance

✅ Spot AI-generated content at a glance

✅ Safeguard who can/cannot use AI for content creation

Learn more about Bynder’s Generative AI today: https://lnkd.in/dv6yde68

 From the community 

• Eric trained Llama 65B on his texts and made a chatbot. Details about his process here. It is fun to chat with.

• LLaMA 2 - Compilation of every resource you need to learn and use Llama 2.

• On holding back the strange AI tide.

 Cool Tools  trending product launches from the last 24 hours

• Taskade - Five AI-powered tools in one, including mind maps and project manager.

• Taylor - Help your engineers train models, without setting up GPUs and deciphering complex libraries.

• Llama API - Easy-to-use API for llama models.

• BlogSEO micro tools - Free AI-powered tools to help you with SEO.

• Layernext - Unified infrastructure for computer vision data.

• Cruderra - Creating docs and architecture as code on one platform.

• Chitchat - A simple LLM chat front-end that makes it easy to find, download, and mess around with LLM models on your local machine.

• Airstrip AI - Legal made easier for every startup.

• Neum AI - Improve your AI's accuracy with up-to-date context.

• Excel formula bot - Write Excel formulas in seconds with AI.

 From the network 

• Llama2.c from Andrej Karpathy - Train a baby Llama 2 model in PyTorch, then inference it with one 500-line file with no dependencies in pure C.

• What comes after SaaS? Bespoke apps for everyone.

• Paddle is starting a 6-week AI launchpad with 30 participants.

 Ben’s Bites News  top posts from the last 24 hours

• Transformers - The Google scientists who pioneered an AI revolution.

• Publishers want billions, not millions, from AI.

• STEVE-1: A generative model for text-to-behavior in Minecraft.

• OpenAI's head of trust and safety steps down, citing work-life balance.

• Linux foundation to refine Ethernet standards for AI and high-performance computing.

• Shopify employee breaks NDA to reveal firm replacing laid off workers with AI.

• NYC subway using AI to track fare evasion.

• FTC thinks generative AI raises competition concerns.

• ‘Judeo-Christian’ roots of Pentagon AI will ensure US military AI is used ethically, general says.

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)