• Bot Eat Brain
  • Posts
  • DeepMind's AlphaFold predicts protein molecules

DeepMind's AlphaFold predicts protein molecules

PLUS: Google Bard says speed > quality

TOGETHER WITH

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

Here’s what’s on the menu today:

  • AI now predicts protein molecules like a boss 🧬 🔥

    Access all of GPT’s functionalities without switching models and more.

  • Apple’s new MacBooks are targeting AI developers 💻 🤖

    Its new M3 chips have neural engines optimized for powerful AI models.

  • Google Bard says speed > quality 🏃‍♀️ 💨

    Google’s chatbot, Bard, now replies in real time.

MAIN COURSE

AI’s protein analysis game = fire 🧬 🔥

In September, we reported on Google DeepMind’s AlphaMissense. It’s an AI tool that looks for changes in DNA that lead to a different amino acid in the protein.

Yesterday, Google DeepMind announced an update for AlphaFold, its AI program that predicts the 3D structure of proteins.

Knowledge is power.

Sounds good, what makes it so great?

It outperforms the traditional methods like “molecular docking” that predict how drugs will interact with proteins in our body.

What did it do before?

When first introduced in 2020, it modeled individual proteins, which was a major accomplishment. Now, it models complex groups of molecules.

AI advances every day, will this have staying power?

This particular advancement could lead to better drug designs, improve the way we edit genes, and more. DeepMind claims it’s a step towards “digital biology,” using computers and AI to simulate and study our molecular makeup.

One small step for protein, one giant leap for humanity. 👩‍🚀

Build an AI ChatGPT for your website in five minutes.

Wonderchat helps you build your own ChatGPT-powered chatbots in 5 minutes.

Wonderchat empowers you to:

🚀 Build chatbots trained on website links and PDF files in 5 minutes.

🚀 Automate up to 70% of your support queries from your site visitors.

🚀 Multilingual chatbots that provide 24/7 round-the-clock support in over 80 languages.

🚀 A chatbot that integrates with over 5000 apps such as Slack, Microsoft Teams, and more via Zapier 👇

BUZZWORD OF THE DAY

Neural Engine

A collection of specialized computational cores on Apple’s Silicon chips.

They’re designed to execute AI/ML functions quickly and efficiently, with the action happening directly on the chip itself.

SIDE SALAD

Apple’s new M3 chips for MacBooks 💻 🤖

In July, we covered Apple’s chatbot, Apple GPT. It was built on a custom LLM framework called Ajax.

Two days ago, Apple announced its new M3, M3 Pro, and M3 Max chips. They’re available for its new iMac and MacBook Pro, available on November 11.

How is this AI-related?

These chips come with an improved neural engine that is specifically designed to accelerate powerful AI models.

The neural engine operates AI/ML workflows up to 60% faster than the M1 chips.

So if I’m not running AI models, it’s useless?

It also uses AI to enhance the performance of the new MacBooks and iMacs. AI tools like noise reduction and super-resolution benefit from the enhanced speed.

When edit videos, features like scene edit detection in Adobe Premiere and Smart Conform in Final Cut Pro get a performance boost as well.

So, they’re not just saying “AI” to boost sales?

The chips unlock workflows previously not possible on a laptop, such as working with large AI transformer models. The 128GB unified memory enables AI developers to handle even larger transformer models with billions of parameters.

A LITTLE SOMETHING EXTRA

Bard got with the times 🏃‍♀️ 💨

Two days ago, we reported on Google’s new partnership. It invested $2 billion into Anthropic.

That same day, Google announced a new update for Bard. The chatbot now replies in real time.

What does that mean?

Previously, it only sent completed responses. Now, it gives you a glimpse of your answer as Bard generates it.

This also means you can cut Bard off if generating something that is less than helpful to you.

I don’t like change.

You can switch between “Respond in real time” or “Respond when complete.”

How do I shut it up?

If it’s babbling enraging nonsense, as chatbots often do, you can cut Bard off mid-sentence by hitting the “Skip response” button.

MEMES FOR DESSERT

YOUR DAILY MUNCH

Think Pieces

How do venture capitalists invest in AI? A former Google employee explains how he invests in LLMs, drug discovery AI systems, and more.

AI is helping, not hurting, the job market. A look at how AI is helping humans be more productive.

The UK’s AI summit. What to expect from Wednesday and Thursday’s event.

Startup News

Meta and Dell partner up. They claim the goal is to streamline the deployment of LLaMA 2 for businesses.

Stability AI, Midjourney, and DeviantArt are in court. A federal judge allowed one lawsuit against Stability AI for copyright infringement to proceed.

Microsoft and Siemens partner up. They introduced an AI assistant to upgrade human-machine collaboration and plan to create more.

Research

RedPajama-Data-V2 — an open 30 trillion token dataset for training LLMs.

Hi-NOLIN — an open-source English-Hindi bilingual AI model.

Hugging Face’s RLHF — It reproduced OpenAI’s first RLHF repository which was first introduced in 2019.

Tools

misgif — put your own face in a popular GIF for free.

GenVid — an AI tool to generate demo videos for your business

Embedding Battleground — compare different embedding models.

Llama Index Chat — Share and create chatbots with your data.

If you like Bot Eat Brain there’s a good chance you’ll like this newsletter too:

👨 The Average Joe — Market insights, trends, and analysis to help you become a better investor. We like their easy-to-read articles that cut right to the meaty bits.

TWEET OF THE DAY

As tempting as it may be to scream in all caps at your LLM, you might get better results by asking it nicely.

Tag us on Twitter @BotEatBrain for a chance to be featured here tomorrow.

AI ART-SHOW

Until next time 🤖😋🧠

What'd you think of today's newsletter?

Login or Subscribe to participate in polls.