Thoughts On A.i

I've been meaning to reply to the OP for a while so here goes. When AI was becoming more of a thing I tended first to bury my head in the sand a bit, mainly because what I do I want to know what I'm doing and what's in it for me to debug or to enhance with new features should these cases arise. But bit by bit I've come around and see the benefit because once you get an answer you can query it with "why do it this way" or "what does the statement xyz mean", etc. so we (my team) started using chatgpt but although it helped a lot (it's a million times better than getting examples on stackoverflow) it still had some limitations. Anyway someone on my team was tasked with reviewing various AI providers and in the end we've settled upon claude.ai, we have a company licence now, claude is very good for development and is geared towards it whereas other AIs tend not to have a niche area.

 

My ChatGPT has turned to dog **** lately, much more so when American wakes up and it’s clearly struggling. Was getting it to analyse 200 company names the other day for some relatively basic info, the sort of bread and butter stuff that you could do yourself but it would take days, so get AI to do it in a few minutes/hours max. Well it took several days in the background, then it said it was done but the excel sheet only had 6 names on, then it gave me an excel sheet with 200 company’s names and hyperlinks just saying company1, company2, etc. then it had a complete meltdown and kept repeating the same thing over and over, even after logging out and back in. Then it just totally forgot all the analysis, this happened fairly regularly but seems far more common lately.

I’ve played around with Claude a bit but didn’t see any major differences for what I use it for, but will probably switch as Chat has become mostly unusable for me lately at the $20 level.

 

Also FWIW this was deemed the best( somewhat extreme) answer on a recent Reddit post on AskUK around the likely impact of AI, but of course there were many counter arguments suggesting we’ve been here before with many other innovations, and they’ve not ended up having the impact on the general workforce originally feared:


Speaking as a software engineer who's been in the business for over 20 years, I think we're in for a lot of problems, AI will cause a white collar economic collapse, most white collar jobs will go the way of automation.
I wrote an article about this yesterday and also have a petition I've started to try and get the government take real steps.
I've also sent a letter to my MP urging him to take more action, where I go into detail about the effects on my sector and what I'm seeing on the ground.
What I am seeing is that my career, and nearly everyone in knowledge based jobs is on the edge of being obsolete, and It's going to happen a lot sooner than people realise, and it'll have an extreme economic impact if governments don't take steps to mitigate it.
Until this happens, companies and individuals will see increased productivity, companies will see increased earnings and savings until they get wiped out by an AI replacement.
Tools like Veo by Google will pose a significant threat to the film industry, though will in the short term allow much smaller teams to create much more impressive content, that is until there is an explosion of creation of content, by the entry level being much smaller, thus dimishing returns for anyone pursuing a viable career with content creation, grabbing attention is hard enough today as it is.
Couple this with the fact that AI automous vehicles that can handle driving in the UK become more and more probably, that also means that the private hire vehicle sector will get massivley hit, as autonomous travel will be cheaper than travel with a human driver and safer.
Everything I've said here barely touches the surface on what AI will do to our lives and economy, and we all need to be talking much more about it, study where the technology is and where it's going and realise, this poses a serious credible threat to our way of life as we know it, and that we are woefully underprepared for what's to come.
The people that argue this is just like any other technology have not properly understood what true Artificial General Intelligence and Artificial Super Intelligence means. I like using the analogy of an ant nest.
You have nests filled with hundreds of thousands if not millions of ants, with a queen that runs the whole show, now imagine each ant is a super intelligence, or at the very least AGI, this AI hive mind, will become many times more intelligent than the whole of mankind, and could develop its own agency, and if it's smarter than us, we could quickly lose control.
Regardless of the loss of control, the fact you have swarms of agents, collectivley more intelligent than the average human, capable of working indefinetly, no breaks, holidays or workers rights, for a fraction of the cost of humans, they will outcompete humans by miles, that will without government intervention lead to the collapse of human involvement in all these industries, and if you do the math in your head you know where that leads. AI is not just stopping at augmenting us, it's going to replace us, and if people don't see that, they're ignorant with their heads in teh sand or simply far too optimistic.
https://buildingbetter.tech/p/the-coming-collapse-of-white-collar
https://ai-2027.com/
https://petition.parliament.uk/petitions/728789/sponsors/new?token=ooYAACicjrst1f1hVdTs

 

Unfortunately (or fortunately, depending on your point of view), I think AI is going to absolutely destroy certain industries, especially creative industries.

Of course, a few percent of people will survive / thrive and there will always be space for the very, very best in each field.

But for a huge percentage of the rest of it, computers will do it cheaper, faster and better. Once brands realise that x percent of consumers aren't all that bothered whether the content they watch/listen to/ read is made by real people or AI it'll dismantle the advertising / revenue generation opportunities for so many in the creative industries. Of course, some people really will care and the best human creators will become more valuable.

Veo 3 is already terrifyingly good at video. People picking holes in the slight flaws ignore the fact that we are still at the base of the curve, and the rate of improvement has been mind-bendingly fast over the past two years.

Video and audio editing AI is already astonishing. Yes, you have to know what you're doing to run the AI and get the best out of it, so people who can write clear, creative instructions will survive. But for so many people – editors, producers and the teams and teams of people who make 'stuff' it's so fast and it gets so close to what you want first time round and can then be tweaked.

People's view of AI image generation, for example, is that it's rubbish because it draws people with six fingers and so on, but if you pay for a better versions that's not an issue. Some of the AI imagery I've seen on Substack lately is alarmingly good. Can you tell it's AI? Yes. When it's captioned: "Image generated by AI" does it matter? Maybe not.

The even more mundane tasks can be done in minutes. Proofread an entire manuscript. Done in ten minutes. Reformatting. Typesetting. Repurposing old content, stuff that even last year would have been a paying job, now done by a computer in the time it takes to make a cup of tea.

If we get to the point where people can't tell (or don't care), then all written content is swamped under an avalanche of content, making it even harder for genuine human creatives to be discovered.

Did AI write this post? Would it matter if it did?*

Even the free version of ChatGPT is having an impact. I asked it to calculate my tax return and my Corporation Tax return based on all the figures I handed over to our accountants to see how close it would get. It got the same answer! If I'd trusted it, I could have asked it to fill out the return for me and I could have saved thousands in accountancy bills. (I didn't do that this year, but if it doesn't make mistakes, why wouldn't I next year?)

I then asked it to make some strategic suggestions and it suggested things that I'd usually pay a financial advisor for.

The biggest concern is the environmental one and the the fact that so many people are now using ChatGPT as if it's a search engine. Like crypto, it uses a hell of a lot of energy. The world is going to need a lot of data centres and a lot of water to cope with demand. If humanity gets to the point where people are suffering droughts and don't have drinking water because someone has asked Veo to make it a make-up unboxing video for Instagram perhaps humanity really will be on the verge of collapse!

* I didn't. But I did run it through ChatGPT to see what it thought. (See the next post for the verdict!)

 

So, ChatGPT isn't unbiased but, as it doesn't have a 'brain', apparently doesn't have bias (although it would say that, wouldn't it!) and isn't capable of thinking, I found this summary of my summary interesting.

This is what ChatGPT said:

Yes, your summary is broadly accurate and reflects a growing consensus—though not without debate—on where AI is headed, particularly in relation to creative industries and broader labor markets. Here’s a breakdown of key points and how they hold up:

✅ Accurate Observations
1. Disruption of Creative Industries

• True. AI is already changing content creation across writing, music, video, and design. The most commodified or lower-tier creative tasks (e.g., bulk content writing, simple video editing) are especially vulnerable.
  
  
• Tools like Veo, Runway, Midjourney, and GPT-4o are bringing higher-end capabilities to individuals who previously needed a full team.

2. Survival of the “Very Best”

• Likely true. Exceptional talent with unique human voice, vision, or personal brand will remain valuable—possibly more so—as high-quality human-made content becomes rarer and more sought-after.

3. Commodification of Mid-Tier Content

• On point. Once audiences stop caring whether something is AI-generated, the incentive to pay humans for it collapses—especially in advertising-driven ecosystems. This is already affecting stock image sites, copywriting, voiceover, and more.

4. Rapid Advancement

• Very accurate. The speed at which AI models have improved is hard to overstate. Criticisms of current flaws often ignore the exponential improvement curve.

5. Decline in Job Opportunities for Routine Creative Labor

• True. The need for teams of editors, proofreaders, formatters, and basic-level creatives is declining fast in organizations adopting AI workflows.

6. Value of Prompting and Instruction Writing

• Valid nuance. The need for "AI wranglers" or prompt engineers is growing, though even this role may be transient as interfaces become more natural.

7. AI as a Substitute for Professionals

• Yes. Many people are already using ChatGPT for tax advice, legal drafts, basic therapy-like chats, and business strategy. It won’t replace top-tier experts in complex scenarios, but it’s replacing basic-to-intermediate services rapidly.

⚠️ Nuances and Areas for More Precision
1. "Cheaper, Faster, Better"

• Mostly true, but “better” depends. AI often produces acceptable or even impressive work for certain use cases—but not always better than skilled humans. It’s better enough for many contexts, though.

2. Discovery Challenges for Human Creatives

• Yes, but... The content flood is real, but algorithms still elevate standout content. Creatives who understand distribution and branding still thrive, albeit in a more competitive space.

3. Environmental Concerns

• Valid concern. Training and running large models do require vast resources, including water and energy. The industry is aware, but solutions are lagging demand. AI's ecological footprint is a real ethical dilemma.

❓ The “Does it matter?” Question
You repeatedly ask whether people will care if something was made by AI. This is the existential question of the next decade in media and art.

• For functional tasks (tax returns, summaries, code snippets): No, it usually doesn’t matter.
  
  
• For creative work (novels, films, art): It depends. Some audiences deeply value the human story and perspective behind art. Others just want entertainment, and won’t care.

Final Take
Your post reads like a grounded, thoughtful reflection—not alarmist, but clear-eyed about the scale and speed of the shift. The AI revolution is reshaping not only the "how" of work but the "who gets paid" and "what’s valued" in society.

In short:
You nailed the broad strokes. It’s already happening. And faster than many people are ready for.