- Generative AI - Short & Sweet
- Posts
- 🍓 Friday Tech Updates + Everyone can Code! ✨ closes
🍓 Friday Tech Updates + Everyone can Code! ✨ closes
60 minutes to build a professional offline (secure) AI application from scratch, with the features you need and a professional UI, all without coding experience?
OpenAI’s next moves?
What is this Doom AI game everyone is talking about?
4500 developer years saved in Amazon? How does that make sense?
Read all of this below.
Reading time is 136 sec. Let’s go!
Learn how to prompt AI so that it guides You to build what You want 🫵
Since yesterday, our Everyon can Code!✨ participants can now access all resources on building and packaging (secure) AI-offline applications.
In the right way, you can do it in under 1h, including a beautiful AI with v0.
Version 1.
Using v0 you can quickly with natural language build different styles of your UI.
Version 2.
🚨 In two days, we release a lesson on building a professional RAG system—something companies pay millions to implement.
We let it build from scratch in under 60 min. - plug in the data you need, and connect it with the front end you want.
The AI does it.
Again, in under 60 minutes - let that sink in. Things have changed, many don’t know it yet.
⚠️ Enrollment window1 closes on September 1st permanently!
OpenAI Updates: Strawberry → Orion, and Further Funding ($100B)
New Model training GPT-4o’s successor
OpenAI is preparing to launch a new AI model, codenamed "Strawberry".
This model is designed to solve complex mathematical and programming problems, as well as it will have further reasoning capabilities.
Strawberry is tied to OpenAI's Orion, GPT-4o's successor. Strawberry's key feature is generating high-quality synthetic data, vital for training Orion.
OpenAI showcased Strawberry to U.S. national security, highlighting its impact on security and AI transparency. Expected release: fall, possibly integrating into ChatGPT.
More funding for OpenAI
If you want to have stronger AI models, you need more computing power. Latest investments:
xAI/Elon Musk: Investing $6B in Memphis for a supercomputer with 100,000 Nvidia H100 GPUs, and more.
Meta: Expanding to 600,000 GPUs by 2024, including 340,000 Nvidia H100s, and more.
Google: Allocating $50B in 2024 for AI and cloud, backing initiatives like the Gemini AI model series.
OpenAI knows that of course and so they are in discussions to secure a new funding round that would value the company at over $100 billion.
Thrive Capital is leading this effort with an expected investment of about $1 billion, and Microsoft is also anticipated to participate. This would mark the largest capital infusion for OpenAI since Microsoft's $10 billion investment in January 2023.
Google’s GameNGen creates Doom (video game) in real-time
Google's rolled out an AI system called GameNGen, and it’s doing something pretty wild: simulating the classic game Doom without a traditional game engine.
Yeah, you heard that right—no engine needed.
This is a significant step in AI-powered game development and could shake up the entire gaming industry.
How does it work?
The steps involved:
Neural Engine: Runs on a neural model, using a TPU, generating gameplay at 20+ FPS.
Diffusion Models: Predicts and renders frames using generative AI.
Reinforcement Learning (Not from Human Feedback): Trained on Doom data to simulate game state updates.
Real-Time Play: Enables interactive gameplay with visual quality similar to the original Doom.
Limitations: Only works for Doom, with a memory span of ~3 seconds, affecting long-term stability.
Amazon's CEO says its AI tool has saved a crazy amount of time
(Source)
Amazon CEO Andy Jassy revealed that Amazon Q, the company’s generative AI assistant, saved 4,500 developer-years by slashing software upgrade times. The transition to Java 17, once a 50-day task, now takes just hours with Amazon Q.
You can be at the forefront within your organization, driving this productivity explosion.
(We close the enrollment window end of this week.)
So excited about Everyone can Code!✨
It is the world's most enabling AI course.
Why most enabling?
- You'll learn to build anything (software or feature).
- You'll be able to build your version 1 in under 1 hour.
- Build anything: web or mobile apps, browser extensions, AI applications (from secure offline SLM to RAG with LLMs!), and all kinds of frontends are needed.
- Most powerful: have access to THE Exclusive Community of AI Builders (Discord) 🤝 We build projects together.
A new product development era starts: Build fast (e.g. 3 features), get real user reaction (e.g. pixel), kill all features but the winner. Rinse. Repeat.
If you are unsure about taking the course or not, I am happy to give you a call and we see together if it is for you. → [email protected]
That’s a wrap.
Have a good weekend!
-Martin
Everyone can code! ✨ —> Landing page.
My upcoming training is on GenAI, AI Agents, and AGI.
Would you like to sponsor a post? —> www.passionfroot.me/ai
Spread the word! Referral program.
My book: https://a.co/d/eMosWDc
Our webpage: https://generativeai.net
1 Enrollment window for batch 1.