TOGETHER WITH

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

Here’s what’s on the menu today:

• Remake Doom with AI 🎮 🔥

  Google researchers unveiled GameNGen.

• Amazon and Anthropic lock lips 👄 💕

  Alexa will reportedly be powered by Claude 3.5 Sonnet

• In the mind of a newsletter writer 🧠 👨‍💻

  I did a podcast about creating a newsletter.

  New here? Subscribe!😎

Peep today's ‘What Would You Do?’ at the bottom. 👇

MAIN COURSE

Create your own games with AI 🎮 🔥

On Tuesday, Google researchers unveiled GameNGen. It’s an AI system that simulates the classic shooter Doom without using a traditional game engine.

What's the big deal?

This marks the first time an AI has fully simulated a complex video game with high-quality graphics and interactivity in real-time.

How does it work?

GameNGen uses a neural network to generate gameplay at 20 frames per second. It runs on a single Tensor Processing Unit (TPU), Google's custom AI chip. The system uses a diffusion model to predict each frame.

Why should I care?

This could revolutionize the $200 billion gaming industry. It may lead to faster, cheaper game development. AI-driven engines could create new genres of dynamic, evolving games.

What are the limitations?

Currently, it’s tailored specifically for Doom. More graphically intensive games would likely require more computing power, and developing a general-purpose AI game engine would require a different approach.

What's next?

This technology could have far-reaching implications for virtual and augmented reality, autonomous vehicle training, smart city simulations, and more.

The bottom line:

GameNGen represents a significant step towards AI-powered game creation and simulation, potentially transforming not just gaming, but various industries relying on complex, real-time simulations.

Click here to check out the paper.

SPONSORED BY LLM BEEFER UPPER

Instantly improve your LLM outputs

The LLM Beefer Upper automates chain of thought, using multi-agent prompt templates.

97% of users say the final post-reflection result is better than the first. Choose from popular pre-built templates or design your own, for whatever LLM process you need.

New users get 50 credits.

SIDE SALAD

Amazon and Anthropic lock lips 👄 💕

Amazon is reportedly powering its “Remarkable Alexa” with Anthropic's Claude AI.

What's new?

• Amazon is revamping Alexa with more advanced AI capabilities.

• The new version will use Anthropic's Claude 3.5 model.

• It's expected to launch later this year as a subscription service.

Why the change?

• Amazon's in-house AI reportedly faced major latency issues.

• Early versions took 6-7 seconds to respond to prompts.

• Claude's superior performance pushed Amazon to partner up.

What's the deal?

• Amazon invested up to $4 billion in Anthropic last year.

• Anthropic will use AWS for computing infrastructure.

• UK regulators are investigating the deal's impact on competition.

How much will it cost?

• Amazon plans to charge $5-$10 per month for “Remarkable Alexa”.

• A free tier with “classic” Alexa will remain available.

• Some employees are skeptical about users' willingness to pay.

Why should I care?

• This move mirrors Apple's integration of ChatGPT into Siri.

• It shows big tech's shift towards hybrid AI models for voice assistants.

• 2024 is crucial for Amazon to prove Alexa's revenue potential.

What's the catch?

• Users who've already invested in Alexa devices might resist additional fees.

• The subscription model faces skepticism internally and externally.

• Apple is reportedly developing a competing tabletop robot device.

Click here for more on the alleged partnership.

SWEET DESSERT

In the mind of a newsletter writer 🧠 👨‍💻

On Friday, I had an enlightening chat with Mythri about running a successful AI newsletter. Here are the top insights I shared for aspiring newsletter creators:

Data is King

• I check analytics daily, focusing on engagement metrics over raw subscriber count.

• I maintain a 40% open rate by regularly purging inactive subscribers.

• I analyze data every 30 days to avoid emotional decision-making.

Quality Over Quantity

• I turned down a $150,000 buyout offer, believing in the long-term potential of AI content.

• I prioritize high-quality, engaged subscribers over inflated numbers.

• I don't chase vanity metrics at the expense of true audience value.

Monetization Matters

• Current model: Sponsorships (high demand allows daily newsletters).

• Future goal: Building a premium community for networking with AI executives.

• Advice: View your newsletter as a business from day one.

Content is Key

• I spend 4–5 hours daily writing the newsletter.

• Our focus is on breaking down complex AI news into actionable insights.

• Strategy: Make content simple, relatable, and immediately useful to readers.

Unexpected Skills Learned

• Time management became crucial due to the newsletter's demands.

• Learning to say “no” to misaligned sponsorships, even if it means leaving money on the table.

• Insight: Protecting your brand and audience trust is more valuable than short-term gains.

Bonus Tip: I recommend new creators write their own content initially to gather data and find their voice before considering outsourcing.

To watch the podcast, click here.

YOUR DAILY MUNCH

Cool Tools ⚒️

PodSnacks

An AI-powered podcast summarization tool.

Streamline Your Podcast Experience

Receive bite-sized, AI-generated summaries of your favorite podcasts delivered directly to your inbox and save dozens of hours of listening time.

Think Pieces 🧠

Will AI ever produce true art? Ted Chiang, the science fiction writer, claims its potential is theoretical, and its success has been in lowering our expectations.

Startup News 💰

DeepMind and UC Berkeley enhance LLMs’ performance. They optimized the use of compute resources during inference rather than scaling up model size.

Research 👨‍🔬

DisTrO — reduces communication between GPUs during AI training by up to 10,000 times, enabling the training of large language models.

Idefics3-8B — a guide to developing vision-language models addressing current state-of-the-art methods, key challenges, and potential research directions.

FURRY FRIENDS

Guess his name? Email your guess. 🐈

Respond to this email with your pet’s name and pic for a chance to be featured here tomorrow!

MEMES FOR DESSERT

AI ART SHOW

WHAT WOULD YOU DO?

Housing prices are too high… 😳

Pick your kitty condo? 👇

Ideas? Comments? Complaints?e

Respond to this email or hit me up on 𝕏.

Until next time 🤖😋🧠