Trends in AI: Training Costs and Diffusion Speed
dewierwan.github.io · 176 words · saved by 1 readers
Interactive visualization of AI training costs and diffusion speed trends over time
Explore trends in AI training costs and diffusion speed through the combination of exponentially increasing frontier model costs and rapidly decreasing replication costs for existing capabilities. 📈 Frontier Growth Training a state-of-the-art AI model requires exponentially more computational resources each year. 📉 Capability Diffusion Meanwhile, achieving any given capability becomes exponentially cheaper through hardware improvements, algorithmic progress and more high-quality data. Frontier Model Training Trends Resources required to train the most capable AI models over time…
saved by
related reading
- Scaling: The State of Play in AIoneusefulthing.org
- AI Model & API Providers Analysis | Artificial Analysisartificialanalysis.ai
- There's An AI For That® — The front page of AItheresanaiforthat.com
- How much does it cost to train frontier AI models? | Epoch AIepoch.ai
- Datacurve | The data engine for frontier AIdatacurve.ai
- My picture of the present in AI — LessWronglesswrong.com
- AI progress is about to speed up | Epoch AIepoch.ai
- Together AI | The AI Native Cloudtogether.ai
- CAIS AI Dashboarddashboard.safe.ai
- Modal: High-performance AI infrastructuremodal.com
- Parsed | Custom, interpretable AI systems that continuously learnparsed.com
- Small Models Have Arrivedcalv.info