Post

AI has impacted me significantly, personally. It is coming for my job first, I'm a Software Engineer. So I am no AI fanboy. But honestly? I can't wait for it to destroy certain companies. Adobe first, one of the most evil companies in all of software and NONE of their products are safe.

49 reposts470 likes

Been reading student essays for a scholarship. 11 of the 13 were AI generated, 1 was plagiarized, and one applicant didn't include an essay. The statistical odds of the 11, from 6 different high schools, having the same talking points is impossible without AI. Disqualified. Stupidity at its best.

I do not believe we have to accept AI replacing us. 1) Don't contribute to AI by taking jobs that train AI. 2) Don't take jobs "identifying workflow items that could be replaced by AI" 3) Stop using any AI that generate revenue for its owner. 4) Rise up in revolution.

The sad thing is I feel like AI COULD have been a useful tool if every single corporation in the country hadn't immediately been like "how can we use this to avoid paying people for their work?"

And then got in front of cameras every chance they got and screamed “we just eliminated white collar jobs boys!!! Wooooo!!!!”

I hope it so thoroughly crushes every single company that has utilized it to eliminate employees that no amount of venture capital can resurrect them. That eventually one day a company exec even absently saying "AI" gets roasted to the point they have to move to Russia, never to be heard from again

Salesforce experience? Chime in. AI is like Salesforce, do it my way, it's easy, your way will cost you. CRM coding morphed from coder to coding added value. Do this by UNDERSTANDING the business needs and offering unique insights YOU saw elsewhere. You know real-world pitfalls and opportunities.

AI is ultimately going to burst as a bubble when govts realise they can’t afford to have 20% plus of its highest paid workforce unemployed plus another 20% of semi skilled employees. When that happens AI will be regulated out of existence.

The problem with this is always the same though: it only takes a small portion of nations NOT to do this and they inherit all the advantage.

Yeah, you’re right, but it will be much easier for some individual to offer much cheaper feature-parity alternatives, if people want to get off the Adobe endless subscription train. That’s my only point

And Oracle. Larry Ellison is a real POS. And hopefully M$ too. Big Brother incarnate. Am SO glad I'm out of that game after 48 years. There was a time when there were few of us and we were treated well - a long time ago.

My optimistic take is that one day all the CEOs will be replaced by AIs that do their jobs better & more reliably than they ever could, and without the desire for yachts & planes & mansions the millions stolen from their employees every day can be redistributed. Unlikely, but I'm allowed to dream.

Yeah I often have some form of this optimism about AI ushering in a period of information stability just because people ask AI questions instead of Google now and we’ve seen even Elon can’t bias Grok without wrecking it; so most of the time Grok calls out Elon and Trumps BS

So there’s a version (probably unlikely) of the next few years where people can ask an ai on each social platform of some piece Of propaganda they are looking at is true and it slowly discredits all the bullshit cannons on the right

My experience of AI as a software engineer with decades of experience has been mixed. Yes, it can do amazing things quickly but soon unravels once there's an issue with what it produced or it doesn't quite work as expected, and often it is just plain wrong and will not be told otherwise.

This is where agentic swarms and iterative passes really shine When you have 48 specialized agents all with their own expertise and hierarchy you can blast through anything it creates that is buggy or incorrect and even find bugs you didn't know existed And it'll only get stronger with new models

48+ agents with increasingly aggressive, competing priorities, specialties and goals... what could possibly go wrong? Before you know it they will have developed their own internal usenet forum and be arguing about how insane the other ones are and my god did you see what shit number 5 implemented.

You actually end up more token efficient because the jobs are done in small parts by specialized agents with very efficient and optimized prompts They take less time to complete, less reasoning and parsing involved, and result in less bugs that use up tokens for QA/fixing

I used #Corel #PhotoPaint 9 for a decade because #Photoshop remained absurdly high for that time. I finally upgraded a decade ago to #CorelPhotoPaint 15, which was still far more affordable and didn't require a subscription.

Good question, but I think if wrecking their products was going to hurt Adobe they should have collapsed a long time ago. It’s more the former

Though I haven't found the interview, one of the main creators of AI spoke about it not being near ready to do all the things they claim or want. He posited that it needs years more testing and review before it's ready, so you're probably spot on about how much things will be screwed up.