Category: Sculpting

RLHF – reinforcement learning through human feedback. This fanciful and seemingly innocuous phrase used in GenAI to moderate the output it generates hides a dark truth. GenAI needs content-moderation on a massive scale, because its inputs are the ‘vast corpus of humanity’, making it impossible to control. Some writers glibly talk about how such moderation, practised through the use of third parties, bring about ‘ethically aligned’ output, completely ignorant, wilfully or otherwise, of a vast environment of clearly unethical practices. We owe this revelation to an old institution: good journalism.

Neither panic nor euphoria is an answer to technological changes but simple understanding of its implications, its ramifications, both in breadth and depth. Neither gaping in awe nor recoiling in fear is a sensible response. Since, despite all the hype, it is reasonably clear that AI is here to stay, wisdom lies in drawing out opportunities for various businesses. Incidentally, this holds important lessons for education too.

We live in a culture which celebrates quantity against quality and is indifferent to origin. The desire for popularity, mass acceptance, career mobility and prestige and the hope of a short public memory forces people to find ‘short cuts’. An entire industry of ‘research papers’ literally opens the path to plagiarism.

Trillions of dollars in business are an alluring prospect in the AI products/platforms space, which will attract billions in investment. Throwing money is not going to fetch results. Just as it happened in the early stages of telecommunications, which was a huge learning for everyone including and especially the companies and governments – because everybody got it wrong – the AI market too will force interested parties to think carefully before plunging. Clinging to entrenched notions could be a path to disaster.

While there is tremendous enthusiasm for the use of AI, the question for makers of AI is when it will begin to fetch profits, which begs another question – when will users be ready in large numbers to pay for the use of AI agents/products?