Raw LLM Responses

Well as a techie working on AI tools on job this is my honest take on it. 1. Its freaking good at basics. But sucks at complex or scenario specific tasks. 2. Yes we can train AI for specific tasks.. but any little variable makes it go haywire.. so 'someone' that isn't AI needs to 'maintain' it. 3. AI Maintenance/Updates are both hardware and software wise. 4. Its a productivity tool, not a replacement of humans! I cant imagine not using using google and instead searching manuals in paper format. AI further improves that. 5. But it also risks cognitive decline if misused. Just as schools don't allow google in exams we need regulations on AI to ensure the remaining workforce adopts to it. 6. We will be in a future where rote memorizing stuff is obsolete! To be in game, you need to grasp fundamental and AI can help mastering fundamentals. 7. Will it cause job displacement? not as much as being hyped. Did we lose horse drivers, horse farms, Carriage makers and related jobs? yes.. did it also create a much later set of new jobs? Yes and that would also be the case with AI 8. The only problem is the speed of roll out. Job loss is exponentially greater than new jobs being created and human greed, not AI, will cause downfall if blindly chased after. So the issue is crony capitalism not AI. 9. Any organization hoping to 'automate' 80-90% or even 100% aka replace humans will have a rude awakening. 10. Its a glorified pattern recognition program which is trained on a ridiculous amount data from across the globe and our own data in the name of personalization. It has no judgment or rationale. Its doesn't even 'understand' the words its spews out! But designed to be confidently wrong anyways! So all the AI going rogue and what not hype is a marketing gimmick that is trying to convince everyone it has Intelligence, but it doesn't. Like really! Its as intelligent as notepad! does perfectly what its programmed for.. in this case pattern recognition based on existing data and predicting next 'token'. So it needs humans to train it and often its easier to train a human in most scenarios, if we are looking at overall cost efficiency (Yes including the datacenters, the power they consume, the amount of engineers you need to maintain and update it and by extension the subscription you are paying for the said AI). And lets not bother using human errors as an excuse, we all have heard of AI hallucinations at this point.