Not yet on that front, rather voice actors, artists producing digital assets on games, iconography for applications, decorating motives for clothing and furniture.
On IT industry specially, I think LLMs generating code are going to become as transformative as compilers were to Assembly developers.
Yes, currently they still go through "generate language XYZ" as intermediary step, however they eventually will improve to the point, that intermediary step will no longer be needed.
And like the Assembly developers that deemed to manually assess the machine code generated by compilers and scoffed at the generated output, eventually had to cope with optimising compilers turning their knowledge into a niche field, same will happen with common programming languages.
Yeah sure, there will be some Compiler Explorer kind of way to see how the AI maps its decisions to e.g. RISC-V Assembly, but only those in the know will bother looking into it, either by curiosity or need.
Some current examples ongoing trends, more specifically IT, in enterprise consulting, we have evolved to less coding, more plumbing, where existing SaaS products get integrated with each other, mostly via configuration, or integration scripts.
Well turns out, some of those common patterns can be used to teach AIs, and let them automate integrations, instead of manually write them by hand.
Not yet, those optimising compilers also took a couple of decades to get right, still miss some stuff like auto-vectorization, however the job of Assembly programmer is for all practical purposes gone.
The assembly programmers gave way to many more higher level programmers, as it became possible for more people to program, and more kinds of programs to be written.
On IT industry specially, I think LLMs generating code are going to become as transformative as compilers were to Assembly developers.
Yes, currently they still go through "generate language XYZ" as intermediary step, however they eventually will improve to the point, that intermediary step will no longer be needed.
And like the Assembly developers that deemed to manually assess the machine code generated by compilers and scoffed at the generated output, eventually had to cope with optimising compilers turning their knowledge into a niche field, same will happen with common programming languages.
Yeah sure, there will be some Compiler Explorer kind of way to see how the AI maps its decisions to e.g. RISC-V Assembly, but only those in the know will bother looking into it, either by curiosity or need.
Some current examples ongoing trends, more specifically IT, in enterprise consulting, we have evolved to less coding, more plumbing, where existing SaaS products get integrated with each other, mostly via configuration, or integration scripts.
Well turns out, some of those common patterns can be used to teach AIs, and let them automate integrations, instead of manually write them by hand.
https://blog.hoyack.com/top-7-tools-for-effective-agent-orch...
Do they work as well as the sales pitch?
Not yet, those optimising compilers also took a couple of decades to get right, still miss some stuff like auto-vectorization, however the job of Assembly programmer is for all practical purposes gone.