I keep telling VCs and Angel investors that now is the time to look for relatively under-valued high growth startups using Reinforcement Learning to solve real world B2B / engineering / logistics problems ...
The balanced NPU/GPU/CPU with close cache ram, such as the recent Lunar Lake chip, coupled with better integrated GPUs in consumer class laptops, and a nice web GPU webgl api .. make for a lot of capability for problems that are well solved by Monte Carlo simulation ... and RL generally.
3D apps are now really doable in browser on current devices... and the browser is a great delivery device for applications, avoiding platform specific installs and dependency hell.
The balanced NPU/GPU/CPU with close cache ram, such as the recent Lunar Lake chip, coupled with better integrated GPUs in consumer class laptops, and a nice web GPU webgl api .. make for a lot of capability for problems that are well solved by Monte Carlo simulation ... and RL generally.
3D apps are now really doable in browser on current devices... and the browser is a great delivery device for applications, avoiding platform specific installs and dependency hell.