I’ve been in tech for nearly 20 years and I’ve immersed myself into almost all the hyped up stuff, from 3D printing back in the day to AR/VR to crypto to AI.
AI is the only thing that feels like the real thing. It’s already being used in my work extensively, from coding to content generation.
If you’ve found yourself underwhelmed by AI, you’re either not using the latest and greatest stuff, or your problems are too complex for AI to solve - for now.
You're exactly right, and I still find myself shocked every week by how little people are making use of it.
I think part of the reason is the same as another phenomenon I've seen before. Most people don't want to automate their work. In fact, they don't even want to change their behaviour at all. They have tasks which they are 100% confident and comfortable in, and completing them gives them satisfaction. As long as they can do that and keep their job, they will continue in that manner.
As a business owner, it makes me want to pull my hair out. There is so damn much wasted time and energy in this world. But I also can't blame them. They're happy, their boss is happy, the company is happy (for now).
> As a business owner, it makes me want to pull my hair out.
My thought is that it's long since an employee got rewarded based on output or their work done. They (we) get rewarded for having their butt in their seat for 8 hours a day, no matter how they spend that time.
If me using AI tools to do my work more efficiently actually meant that I could go home from work two hours earlier, I would be all over it. If using it however means I need to do more stuff and the owner reaps all the rewards.. I'm not gonna bother.
> They (we) get rewarded for having their butt in their seat for 8 hours a day, no matter how they spend that time
Just depends on where you work.
Doesn't part of you want to make the most of your time? Is there nothing more enjoyable or beneficiary you could fill the extra time with, even if it's still "work"?
I don't agree you'd be doing "more stuff", it's about getting the same value/productivity out of the same or less input
I run a small D2C brand on the side that's been lagging behind because I need to create fresh creative collateral for it and I've never had to time to get new set of pictures taken.
The Midjourney V6 > Magnific pipeline helped me create enough pictures in 3-4 hours of prompting to keep the marketing pipeline filled for the next year. And the quality is so, so good that most of my agency friends say only a handful of people in the state can create anything remotely similar.
People will be shocked at the quality of creatives even small brands will be able to put out in the next couple of years. You can have Versace-quality luxury product shots for $20/month in subscriptions.
Yes, I've been surprised by this as well. I expected that an older person who had done the same job for 10 years would be like this. But I'm seeing tech oriented people not want to use it.
> or your problems are too complex for AI to solve - for now.
i am not very interesting in the current AI wave, but I gave chatgpt a try and asked them to generate some sample code using a specific library.
Now, i had read the library docs top to bottom, and the generated code was essentially wrong.
The thing that got me, i could definitely think of the specific manpages (or webpages) and samples chatgpt had used/shuffled to generate the code.
Needless to say, it shows how it can do a patchwork of text without really understand what's going on: the various pieces of code came from some different sources, were "stitched" together pretty much correctly, but the code was essentially doing the wrong thing.
Don't let perfection be the enemy of the good. Today ChatGPT might only get you 80% of the way there, but you can still fill in the 20% and still come out ahead.
For many of my tasks it's getting me 90% of the way there because there's simply many more code examples and much more documentation in my domain. It still makes mistakes and it's not great at reasoning, but ChatGPT 4 (forget the free tier, always use 4) is at a level that lets me be more productive than if it had not existed.
It's really fixed mindset vs growth mindset with respect to technology use.
Fixed mindset = this sucks because it doesn't do everything I want. It's dead in the water.
Growth mindset = this doesn't yet do everything I want, but I can still use it. Or maybe I'll patiently wait till it becomes something useful. But I'm not going to dismiss it out of hand.
Yes. Too many people who have only used the free ChatGPT 3.5 come away thinking, "if this is all there is, I've got nothing to worry about". It's not until they pay for ChatGPT 4.0 and use it for a while that they realize they've been dismissing a straw man all this time.
To be fair, ChatGPT 4 isn't available everywhere and $20/mth though cheap to Americans, isn't that affordable to the rest of the world.
But it is a vast improvement over 3.5 and actually a game changer for many things. Even a year after its launch, no other LLM has come close:
These are really important questions. Because if you used ChatGPT-3.5 and prompted it with "write code", it would be akin to giving your kid a 10 year old Intel i3 laptop with 2GB RAM and then being baffled that your kid isn't interested in computers.
If you are reading this and getting upset at the lack of information, remember this is just CNBC. So its more like an introduction to the concept of AI job loss for the target readership. (Which probably isnt as technical or complexity savvy as the average hacker newser)
Granted, I would not have considered this worthy of posting.
Well Keynes was predicting the possibility of 15 hour weeks back in 1930 thanks to technological advancements and a preference for leisure over labour. I recently re-read his paper "Economic Possibilities for our Grandchildren" - https://www.econ.yale.edu/smith/econ116a/keynes1.pdf
He was right, MANY people put in way less than 40 hours of what he would have considered actual work. Scrum Masters, DEI Officers, Management Consultants are not real jobs for the most part. They are mostly in meetings saying vague buzzwords.
For those just reacting to the headline, this article is about leveraging AI in current roles and its effects on employment equality, not job losses due to AI.
When white collar work that can be made remote doesn't spread the wealth because AI makes it obsolete, regulation does. Wealthy countries tend to ban or limit activities at ever-smaller thresholds of risk and that gives lucrative sources of income to less-wealthy and restrictive countries. Recycling, manufacturing, training on privacy sensitive-corpses of data (healthcare data for diagnostics for example, or as we recently saw IP law attacks on Generative AI), fintech, moderation of social media content, etc. The problem with those is that over time the money becomes concentrated on a few major firms instead of spreading around.
That said, with AI, now anyone who has been doing major digital work in non-major markets has an ability to make creative or entrepreneurial projects at a blazing fast speed now. If there are consistent distribution channel inroads into wealthy markets, those entrepreneurs could make a lot. TikTok, Instagram, YouTube, SEO, you name it will get even more competitive.
And in the end the big companies remain the gatekeepers and AI-providers will pull an Uber/Lyft and slowly ratchet up the price of their AI products and hire their profit-maximizing PhDs until they've squeezed every little bit of money out of the market.
The net result is probably the same, concentration at the big companies. Maybe a couple people selling "shovels" and "how I made $5k with 2h of work last week" will get rich along the way though.
What's the point of these statistics without even mentioning the date of their predictions? You could write an article stating that 100% of jobs will be hit, because eventually they will be, but maybe in 200 years?
If we just asked GPT4 this, and it said, "Oh, how many jobs will be affected? About 40% is my estimate" - would we have any more reason to be confident in that than in IMF's estimate?
I'm not trying to look like a doomsayer or not, I just say that I doubt there's any actually useful information in IMF's statement at all.
It's not something you can act on. It's not something you can even check in retrospect if will be true. It really is so much like just words statistically generated to placate.
AI is the only thing that feels like the real thing. It’s already being used in my work extensively, from coding to content generation.
If you’ve found yourself underwhelmed by AI, you’re either not using the latest and greatest stuff, or your problems are too complex for AI to solve - for now.