- AI Fire
- Posts
- 🤖 Tesla Pays $48/Hour For Robot Trainers
In partnership with
Plus: Run 2e29 calculations to train AI?
Read time: 5 minutes
On the AI race!
While others were still getting ready, Nvidia had already taken off in the AI race. But now, with AMD acquiring a company that creates cutting-edge AI servers, it’s like they’ve just put on rocket boots, ready to speed up and close the gap.
What are on FIRE 🔥
AI REPORT
👨💻 GitHub's 2024 Developer Survey Insights
Source: GitHub
A new GitHub survey of 2,000 developers across four countries highlights the growing use of AI tools in software development.
Key takeaways:
97% of developers have used AI tools, with U.S. companies leading in support at 88%.
90% of U.S. respondents report improved code quality due to AI tools.
60-71% say AI tools help with new programming languages.
99-100% believe AI will enhance code security.
48% of developers in AI-friendly companies find their toolchains simpler.
Why it matters: Learning to use AI now can help developers advance in their careers. But soon, AI will be a regular part of standard development. Knowing how to use AI coding tools will soon be a basic requirement for software development jobs.
PRESENTED BY GROWTH SCHOOL
🦾 Master AI & ChatGPT for FREE in just 3 hours 🤯
1 Million+ people have attended, and are RAVING about this AI Workshop.
Don’t believe us? Attend it for free and see it for yourself.
Highly Recommended: 🚀
Join this 3-hour Power-Packed Masterclass worth $399 for absolutely free and learn 20+ AI tools to become 10x better & faster at what you do
🗓️ Tomorrow | ⏱️ 10 AM EST
In this Masterclass, you’ll learn how to:
🚀 Do quick excel analysis & make AI-powered PPTs
🚀 Build your own personal AI assistant to save 10+ hours
🚀 Become an expert at prompting & learn 20+ AI tools
🚀 Research faster & make your life a lot simpler & more…
TODAY IN AI
AI HIGHLIGHTS
🔥 ElevenLabs' AI-powered text-to-speech app Reader is now globally available, with 32 language support and licensed celebrity voices.
Here are some fun examples:
Judy Garland can help you review exercises as you walk to class.
James Dean can read your detective story to you.
Sir Laurence Olivier can read today’s newsletter—just paste the web link and listen!
🌟 AMD buys server maker ZT Systems for $4.9B to bolster AI capabilities, challenging Nvidia.
Our POV: The competition in AI chips is intensifying, and AMD is stepping up to compete with Nvidia by improving its software and hardware. In the past year, AMD has invested over $1 billion in AI software and research, including a $665 million purchase of Finnish startup Silo AI.
💬 Nvidia ACE lets players interact with GPT-4-powered NPCs through natural language.
💻 AI textbooks aim to personalize learning, but facing backlash from 50K+ concerned parents.
🤖 Tesla hires people to train its Optimus robot via motion capture, up to $48/hr. Here’s the application link.
Would you like to apply for this exciting position? |
DAILY AI FUNDRAISING
Goodfire raised $7M to develop tools that make AI models more understandable and editable. Goodfire aims to bridge the gap between AI research and practical application, making AI safer and more reliable.
AI SOURCES FROM AI FIRE
NEW EMPOWERED AI TOOLS
👉 Coldreach find ready-to-buy leads and sound relevant with AI
📌 Hamming AI automats testing for voice agents
📲 Ellipsis is AI code reviews and bug fixes
🌱 Myko Assistant is a research assistant for web scraping and finding contacts
⬇️ Fyli finds your customers in seconds with AI
AI QUICK HITS
🤖 Ex-Huawei ‘Genius Youth’ launches robots to compete with Tesla’s Optimus (Link)
💡 Bitcoin miners turn to AI and upgrade power centers for post-halving survival (Link)
🧩 MIT researchers show LLMs can solve puzzles beyond their training (Link)
📄 OpenAI strikes content deal with Condé Nast for ChatGPT and SearchGPT (Link)
💸 U.S. invested $335B in AI, triple China’s amount (Link)
AI CHART
Epoch AI explores whether AI can keep growing quickly through 2030. They focus on key things that could affect this growth.
Challenges to Growing AI Training by 2030
Electric Power: Large AI training sessions need a lot of power. Data centers may need up to 5GW, and networks spread across the U.S. could use 2 to 10 times more power, allowing much bigger training runs.
Chip Production: Big training sessions need thousands of GPUs, maybe even tens of millions more. A big increase in chip-making is needed, but this seems doable.
Data Shortage: AI training uses a lot of data. There is 30x more text data online than what is used now, which could support 900x bigger training sessions. Using different kinds of data (e.g., images, videos) could triple this amount.
Latency Wall: As AI models get bigger, training takes longer. Processing more data at once helps, but there are limits. Current latency limits should be okay for now.
By the end of this decade, AI labs could train models much larger than GPT-4, using up to 2e29 FLOP (FLOP is basic math calculations that computers perform, the higher the FLOP count, the more powerful the computer is).
However, this might be hard to achieve because the power grid may struggle to supply enough electricity due to limitations, environmental goals, and political issues.
AI TUTORIAL
Adjust the Image Detail by Using Web Editor at Midjourney
Midjourney has launched a new web editor. It brings together many different image editing features into one easy-to-use interface. More updates are on the way.
Open the image you want to edit and choose the editor.
If you want to change something specific, add that detail to the prompt. For example, if you want to change an iris flower to a rose flower, mention that.
Erase the part of the image you want to replace by selecting that area.
If your detail is near the edge and you need more space, adjust the image size by moving the blur line or resizing from the corner.
Finally, click submit to create the new image.
AI JOBS
We read your emails, comments, and poll replies daily
How would you rate today’s newsletter?Your feedback helps us create the best newsletter possible |
Reviews of the day
 |  |
We love hearing from readers and are always looking for feedback. How are we doing with AI Fire? Is there anything you'd like to see more of or less of? Which aspects of the newsletter (or AI Fire Academy) do you enjoy most?
Hit reply and say Hello – we'd love to hear from you!
Like what you're reading? Forward it to friends, and they can sign up here.
Cheers,
The AI Fire Team
Reply