What I find particularly insidious about this stage of the exploitation cycle is the FOMO/HYPE machine of 3rd party tool makers - that the true value and human effort necessary to make good tools work is lost in the noise. Smart colleagues feel at once hopelessly behind and emboldened with wild expectations on what can be achieved with little strategy.
To many on the inside of AI's growth depend on those wild expectations of users to shore up the length of the exploitation cycle. I am excited for the long list of resources you point to for teams working on democratizing AI.
The technical community and everyone else have two greatly different perspectives on AI. The former sees utility in narrow and constrained use of the "general purpose" LLM technologies across a number of well-defined use cases. The latter - engages in a cargo cult. Majority of adverse outcomes are coming from this cargo cult mentality.
That’s an interesting way to break it down, Meta is the latest in this cargo cult race building super intelligence but have not bothered to define what it is or what it is aimed to achieve beyond what appears to be a fomo driven move by the very company that created FAIR and gave us PyTorch and has been a proponent of open source. My guess is all that will change in their pursuit for super intelligence dominance.
What I find particularly insidious about this stage of the exploitation cycle is the FOMO/HYPE machine of 3rd party tool makers - that the true value and human effort necessary to make good tools work is lost in the noise. Smart colleagues feel at once hopelessly behind and emboldened with wild expectations on what can be achieved with little strategy.
To many on the inside of AI's growth depend on those wild expectations of users to shore up the length of the exploitation cycle. I am excited for the long list of resources you point to for teams working on democratizing AI.
The technical community and everyone else have two greatly different perspectives on AI. The former sees utility in narrow and constrained use of the "general purpose" LLM technologies across a number of well-defined use cases. The latter - engages in a cargo cult. Majority of adverse outcomes are coming from this cargo cult mentality.
That’s an interesting way to break it down, Meta is the latest in this cargo cult race building super intelligence but have not bothered to define what it is or what it is aimed to achieve beyond what appears to be a fomo driven move by the very company that created FAIR and gave us PyTorch and has been a proponent of open source. My guess is all that will change in their pursuit for super intelligence dominance.