***There are possible risks of a failed AI Gold Rush / AI Arms Race.***
The trillion-dollar valuations in AI firms depend upon two things to assure their value:
(1) The ROI of the vendor’s customers must be sufficient to sustain their use.
(2) The AI-generated code their customers deploy must have enterprise-quality.
NB: Neither of these is currently guaranteed or assured.
***Customer’s ROI is what’s required. That’s not yet guaranteed.***
Current levels of vendor revenue are not proof of future value to their customers. Au contraire. Every corporation recognizes the as-yet-unproven potential upsides, and the competitive threats to not adopting these technologies if their competitors are. So, all corporations need to try to adopt and benefit from these technologies, STAT. Failing to try could be a breach of their fiduciary responsibilities (to their stakeholders).
The largest users of these technologies are largely the vendors themselves (OpenAI, Microsoft, Google, Meta, Nvidia …). Of course, they must use all possible technology, and demonstrate the value, in order to entice (lure) customers into using these at scale. Future ROI is not guaranteed and has not yet been demonstrated.
***Enterprise-quality is required for all generated code. That’s far beyond today’s quality.***
Ask any of the LLMs and they will each recite the many significent risks and downside of using AI-generated code, let alone in an enterprise environment. Including, but not limited to:
Flaws—Outdate dependencies—Lack of security intent—Suboptimal code—Maintainability challenges—Copyright infringement—Data leaks—Bias—Compliance———and on and on.
It’s reckless to deploy today’s AI-generated code at scale in enterprise.
***The technologies are not yet proven, nor are their trillion-dollar valuations.***
***There are possible risks of a failed AI Gold Rush / AI Arms Race.***
The trillion-dollar valuations in AI firms depend upon two things to assure their value:
(1) The ROI of the vendor’s customers must be sufficient to sustain their use.
(2) The AI-generated code their customers deploy must have enterprise-quality.
NB: Neither of these is currently guaranteed or assured.
***Customer’s ROI is what’s required. That’s not yet guaranteed.***
Current levels of vendor revenue are not proof of future value to their customers. Au contraire. Every corporation recognizes the as-yet-unproven potential upsides, and the competitive threats to not adopting these technologies if their competitors are. So, all corporations need to try to adopt and benefit from these technologies, STAT. Failing to try could be a breach of their fiduciary responsibilities (to their stakeholders).
The largest users of these technologies are largely the vendors themselves (OpenAI, Microsoft, Google, Meta, Nvidia …). Of course, they must use all possible technology, and demonstrate the value, in order to entice (lure) customers into using these at scale. Future ROI is not guaranteed and has not yet been demonstrated.
***Enterprise-quality is required for all generated code. That’s far beyond today’s quality.***
Ask any of the LLMs and they will each recite the many significent risks and downside of using AI-generated code, let alone in an enterprise environment. Including, but not limited to:
Flaws—Outdate dependencies—Lack of security intent—Suboptimal code—Maintainability challenges—Copyright infringement—Data leaks—Bias—Compliance———and on and on.
It’s reckless to deploy today’s AI-generated code at scale in enterprise.
***The technologies are not yet proven, nor are their trillion-dollar valuations.***