打破了“模型变强则脚手架消亡”的线性思维。模型能力的提升并非消灭了架构设计的价值，而是将其推向了更高复杂度、更具挑战性的新领域。AI工程师的核心竞争力正是持续探索这种前沿的架构组合。

这句话点破 AI 产品形态的根本转变：早期 AI 是「你要去的地方」，现在变成「你已在的地方」。流量统计将越来越失真——最重度的 AI 用户可能完全不出现在 Web 访问数据中。未来 AI 竞争的关键指标，可能不再是独立访问量，而是「嵌入深度」：你有多深入用户的工作流。

useful observation, akin to Lovelock's [[AI begincondities en evolutie 20190715140742]]

David posits questions that are all on the application side, what is the impact of using ai. There are also questions on the design side, how do we shape the tools wrt those concepts. Vgl [[AI begincondities en evolutie 20190715140742]] e.g. diff outcomes if you start from military ai params or civil aviation (much stricter), in ref to [[Novacene by James Lovelock]]

26:30 Brings up progress traps of this new technology

26:48

question How do we shift our (human being's) relationship with the rest of nature

27:00

metaphor - interspecies communications - AI can be compared to a new scientific instrument that extends our ability to see - We may discover that humanity is not the center of the universe

32:54

Question - Dr Doolittle question - Will we be able to talk to the animals? - Wittgenstein said no - Human Umwelt is different from others - but it may very well happen

34:54

species have culture - Marine mammals enact behavior similar to humans

• Unknown unknowns will likely move to known unknowns and to some known knowns

36:29

citizen science bioacoustic projects - audio moth - sound invisible to humans - ultrasonic sound - intrasonic sound - example - Amazonian river turtles have been found to have hundreds of unique vocalizations to call their baby turtles to safety out in the ocean

41:56

ocean habitat for whales - they can communicate across the entire ocean of the earth - They tell of a story of a whale in Bermuda can communicate with a whale in Ireland

43:00

progress trap - AI for interspecies communications - examples - examples - poachers or eco tourism can misuse

44:08

progress trap - AI for interspecies communications - policy

45:16

whale protection technology - Kim Davies - University of New Brunswick - aquatic drones - drones triangulate whales - ships must not get near 1,000 km of whales to avoid collision - Canadian government fines are up to 250,000 dollars for violating

50:35

environmental regulation - overhaul for the next century - instead of - treatment, we now have the data tools for - prevention

56:40 - ecological relationship - pollinators and plants have co-evolved

1:00:26

AI for interspecies communication - example - human cultural evolution controlling evolution of life on earth

for - AI evolution - nightmare scenario - US govt may seize Open AI assets if it arrives at superintelligence

AI evolution - projection - US govt may seize Open AI assets if it arrives at superintelligence - He makes a good point here - If Open AI, or Google achieve superintelligence that is many times more intelligent than any human, - the US government would be fearful that they could be overthrown or that the technology can be stolen and fall into the wrong hands

for - AI evolution - projections for capabilities by 2030

AI evolution - projections for 2030 - AI will be able to do things we cannot even conceive of now because their cognitive capabilities are orders of magnitudes faster than our own - Write billions of lines of code - Absorb every scientific paper ever written and write new ones - Gain the equivalent of billions of human equivalent years of experience

for - stats - AI evolution - equivalent of 100 million human researchers working 24/7

stats - AI evolution - equivalent of 100 million human researchers working 24/7 - By 2027, the industry's aim is to have tens of millions of GPU training clusters, running - millions of copies of automated AI researchers, or the equivalent of - 100 million human AI researchers working 24/7

for - AI evolution - prediction - 2027 - all cognitive jobs can be done by AI

for - stat - AI evolution - prediction 2027 - training time - 6 OOM decrease

stat - AI evolution - prediction 2027 - training time - 6 OOM decrease - today it takes 3 months to train GPT 4 - in 2027, it will take 1 minute - That is, 131,400 minutes vs 1 minute, or - 6 OOM

for - AI evolution - prediction - 2027 - AI agent will replace AI chatbot

for - stats - AI evolution - Math benchmark - 2022 to 2024

stats - AI evolution - Math benchmark - 2022 to 2024 - 50% increase in accuracy over 2 years - inference accuracy improved 1000x or 3 Orders Of Magnitude (OOM)

for - stats - AI - evolution - Math benchmark

stats - AI - evolution - Math benchmark - 2021 - GPT3 scored 5% - 2022 - scored 50% - 2024 - Gemini 1.5 Pro scored 90%

• for: instant evolution, AI - instant evolution, progress trap, progress trap - AI - instant evolution, progress trap - robotics