Some users defended the claim that patents accelerate innovation as historically effective in many areas, while many dismissed it for software and AI where they enable harmful secrecy and slow open progress.
Based on 8 visible X reactions from 3 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@BlackHC @recurseparadox just like openai would have no muon, flash attention, speculative decoding, test time scaling, etc without all the open research see this plot from last years neurips and tell me who is benefitting more https://x.com/michalwols/status/2077802064130760993
@michalwols @BlackHC Tuning deep nets to be this good would have never been possible in any closed team. Even the Manhattan project would fall short. Because so many papers in that era came with code we got to this point. Remember how no one could reproduce Deepmind Atari results for 2 yrs?
@recurseparadox Really? It still works as intended in many areas. Software patents have issues bc of the way trolls can abuse it but patents were historically very effective at the very least
@BlackHC Yeah the amazing patent system thats so prevalent in the present day 😆😆
@BlackHC It's not effective in software because the 20 year buffer it gives you too much in a world where things change every 3 months. There is no moat in ideas, only in momentum
@recurseparadox Why was the patent system invented again? Patents were actually accelerationist in their effect (and intention) So try again? 🙃
Some users defended the claim that patents accelerate innovation as historically effective in many areas, while many dismissed it for software and AI where they enable harmful secrecy and slow open progress.
Based on 8 visible X reactions from 3 accounts; directional sample.
Ask a question below.
Published answers will appear here.