Hey folks, this is Ben’s Bites, your daily order from the AI drive-through. Today’s lead chef is Salesforce, bringing a few items on the menu. Served with cocktails from leaving Google employees and model leaks from Meta.
Let’s get to it.
🤌 Our Picks
Highlights if you've only got 2 minutes
1/ Salesforce announces its generative AI plans
After a series of not-good reasons for being in the news last week, it is now there because of the release of its AI projects. There’s a lot to cover from this news that I’m not even sure this is a single highlight.
Salesforce’s own AI model for generating content across different functions within their ecosystem (sales, marketing, customer service, development etc.). This will also integrate with OpenAI and other models. Currently in closed beta.
A $250M fund aimed at generative AI startups with the fund’s motto being “responsible generative AI.” As of now they’ve invested in 4 companies: You, Anthropic, Cohere and Hearth.
And the biggest of it all: ChatGPT for Slack
Salesforce owns Slack, and the ChatGPT app for Slack is live (with a waitlist). As Noah Weiss, CPO of Slack said, OpenAI + Slack - It’s a tasty combo.
It’s a symbiotic relationship. You know, when both parties benefit off of each other. OpenAI uses slack for its team and has sent over 5 million messages in their Slack workplace. Now. Slack will use OpenAI’s ChatGPT for summarising threads, project-specific question answers, and drafting replies to colleagues.
I’m a mere mortal, and as tired as I am of Google vs Microsoft updates, I still haven’t got over how Google can get so behind in the AI race. Then comes this article from Wall Street Journal.
Warning you, it is long. Not just another restatement of Google saying they care about AI safety; this one goes deep with stories of employees who developed the early versions of Google’s chatbot and left when the product didn’t see the light of the day. It also goes over the timeline to highlight previous releases from them like LaMDA, Meena, etc.
Overall, it’s a good overview of what happened with balanced opinions.
Facebook (Meta) released their LLM called LLaMa about a week ago, which outperformed the GPT-3 models. They said it was open source, then allegations came up of them being deceptive because weight were not public.
Well, doesn’t matter because the model got leaked on 4chan. It was supposed to be available for researchers or governments only for the time being but users are running it on their local GPUs. Meta is not shying away from it and talking with the platforms like Huggingface to avoid public-facing products to be created from this leak.
Ready to take your startup to the next level but struggling to find the right developers? Look no further than Lemon.io. (Sponsor)
Lemon.io provides a trusted source of pre-vetted candidates to make your search easier and faster. They offer tested and verified devs, affordable rates, and a zero-risk replacement guarantee.
Need developers with startup grit? Go to Lemon.io and find your perfect developer or team there.
🛠️ Cool Tools
Product launches, updates and demos
Atris - AI community manager, a bot that your community members can talk to. (link)
Replicate Codex - Search, filter, and sort through a vast database of AI models on Replicate. (link)
ChatGPT for Youtube - Free Chrome extension that provides access to the summary of YouTube videos. (link)
Quotify - AI-powered tool to extract meaningful quotes from books, articles or any text. (link)
Logoscapes - Make your logo incredible with AI. (link)
Twig - AI augmented support agents to ask questions to your documentation. (link)
MapDeduce - Understand any document with the help of AI. (link)
Dailybot - AI-powered workflows directly into your chat. (link)
Gen AI videos - Upload photos of yourself and generate amazing AI videos. (link)
Orchard - create highly capable LLM agents that go beyond simple conversational interfaces. (link)
Whimsical AI for mind maps. (link)
News, podcasts, videos, blogs etc
Diagram's founder Jordan Singer talks about automating design, self-learning, and beating perfectionism. (link)
The AI future of frontend development. (link)
Ahead of AI #6 - Train differently. (link)
Stability AI acquires Init ML, makers of Clipdrop application. (link)
Brex brings groundbreaking tools to finance teams with OpenAI. (link)
First principles on AI scaling - How likely are we to hit a barrier? (link)
Data-driven ideas - Turning LLMs into research-powered idea engines. (link)
Overhaul raises $73M in equity and debt, to tap AI for securing physical supply chains. (link)
Microsoft's computer vision model will generate alt text for Reddit images. (link)
The best way to start an AI project? Don’t think about the models. (link)
Facebook's powerful large language model leaks online. (link)
10 free open-source AI projects on Vercel. (link)
Here’s what it would take to slow or stop AI. (link)
D-ID’s new web app gives a face and voice to OpenAI’s ChatGPT. (link)
A recap of winning demos from an emergency ChatGPT hackathon. (link)
The tech behind Artifact, the newly launched news aggregator from Instagram’s co-founders. (link)
How 16 companies are dominating the world’s Google search results - 2023 Edition. (link)
How-to’s and resources
👋 Too many links?! I created a database for all links mentioned in these emails. Refer 1 friend using this link and I'll send over the link database.
Published research papers
Deep dive into Nvidia’s latest model Prismer. (link)
Diffusion Policy - Visuomotor policy learning via action diffusion. (link)
Zeroth-order optimization meets human feedback - Provable learning via ranking oracles. (link)
Larger language models do in-context learning differently. (link)
The BigScience ROOTS corpus - A 1.6TB composite multilingual dataset. (link)
Speak foreign languages with your own voice - Cross-lingual neural codec language modelling. (link)
Foundation models for decision making - Problems, methods, and opportunities. (link)
Can we scale transformers to predict parameters of diverse ImageNet models? (link)
Short, sponsored links
Machine Learning School - A 9-hour, live cohort focused on training, tuning, deploying, and monitoring machine learning models using AWS SageMaker.
Hearth AI, a startup backed by the Salesforce Gen AI fund, is hiring founding engineers.
Founding Backend Engineer - Learn more »
Founding Machine learning Engineer - Learn more »
System is hiring a Senior Research Engineer for NLP. Learn more »
Have a product, service, job, event, newsletter, app, book, movie, tool, or anything you'd like to share with over 45k subscribers?
Advertise | Job board | Community | Investor + Founders connect
🖼 AI Images of the day
Funny memes and pics from around the web
Send this to 1 AI-curious friend and receive my AI project tracker database! Use this link.
How was today's email?