I work for a contact center software company. There are definitely some valid use cases for agentic ai and LLM’s but most of the stuff being attributed to the latest wave of AI is stuff that’s already been possible for a decade at least. The new tech open up a few good use cases but they are definitely incremental rather than revolutionary.
We have found it useful for summarizing interactions, doing evaluations on interactions that would have no evaluation otherwise, and highlighting and summarizing relevant passages in knowledge articles for people on interactions with customers and for customers themselves via the web. It does have value, just a small fraction of that which would be needed to justify the investment being made and the valuations it’s getting.
That’s not a revolutionary change, though. It’s a useful tool sometimes, nobody denies it, but this financial speculation and forced deployment has meant that the biggest impact it’s had is the ease with which one can spread disinformation and generate slop for monetization. Both of those have undergone a revolution as a result, that was my point.
This is why I think LLMs have made the world worse even if I don’t buy the fear-mongering. Regardless of the costs to train LLMs, the costs of having those tools available are worse than the benefits. The amount of extra work bug bounty programs have had to do thanks to AI is crazy.
Come to think of it, there are black markets for selling computer security vulnerabilities between nation-states. Maybe our governments can start flooding those with AI generated security exploits. That’d be funny.
I work for a contact center software company. There are definitely some valid use cases for agentic ai and LLM’s but most of the stuff being attributed to the latest wave of AI is stuff that’s already been possible for a decade at least. The new tech open up a few good use cases but they are definitely incremental rather than revolutionary.
The only industries revolutionized by this latest generation of LLMs is botfarms and content slop peddling
We have found it useful for summarizing interactions, doing evaluations on interactions that would have no evaluation otherwise, and highlighting and summarizing relevant passages in knowledge articles for people on interactions with customers and for customers themselves via the web. It does have value, just a small fraction of that which would be needed to justify the investment being made and the valuations it’s getting.
That’s not a revolutionary change, though. It’s a useful tool sometimes, nobody denies it, but this financial speculation and forced deployment has meant that the biggest impact it’s had is the ease with which one can spread disinformation and generate slop for monetization. Both of those have undergone a revolution as a result, that was my point.
That’s true. I agree that’s the only area its revolutionized.
This is why I think LLMs have made the world worse even if I don’t buy the fear-mongering. Regardless of the costs to train LLMs, the costs of having those tools available are worse than the benefits. The amount of extra work bug bounty programs have had to do thanks to AI is crazy.
Come to think of it, there are black markets for selling computer security vulnerabilities between nation-states. Maybe our governments can start flooding those with AI generated security exploits. That’d be funny.