A thread for unboxing AI
The rapid progression of AI chatbots made me think that we need a thread devoted to a discussion of the impact that AI i
By now, I'm sure that a lot of people have seen the post from Matt Shumer, co-founder of OthersideAI, that went viral. It's a very lengthy post, but it can be more or less summarized as follows:
I think we're in the "this seems overblown" phase of something much, much bigger than Covid.
We are in an uncomfortable period where a lot of people in the industry are sounding warnings, a lot of people in the industry are pushing back on those warnings, and most people outside the industry don't have enough information to judge who is correct and therefore have to draw inferential conclusions based on their own biases and their assessments of speakers' motives.
Needless to say, a lot of people who work in Hollywood are unnerved by the ability to create this AI fight sequence between Tom Cruise and Brad Pitt with a two-sentence prompt.
Needless to say, a lot of people who work in Hollywood are unnerved by the ability to create this AI fight sequence between Tom Cruise and Brad Pitt with a two-sentence prompt.
It's over. Old heads like us will still want to see non ai stuff but once the newest generation is raised on a healthy dose of ai content it will seem completely normal to them.
It's over. Old heads like us will still want to see non ai stuff but once the newest generation is raised on a healthy dose of ai content it will seem completely normal to them.
Flesh and blood movies stars will always retain rights over their likenesses. The trick for AI entertainment is to acclimate the audience to "stars" that are entirely digital. That will take a bit of time, but I agree that it is nearly inevitable. It's just a question of how long it takes.
Yeah, it's just a matter of conditioning, will take 2 generations or so to be the vast majority but AI's role in every facet of the industry will grow yearly, like basically everyone else's industry.
On the plus side, everyone will have the ability to produce a mission impossible type budget movie at their fingertips and that will hugely put the emphasis on the story and message. Barrier to entry is nothing and people who could have never had the opportunity can now tell their story.
By now, I'm sure that a lot of people have seen the post from Matt Shumer, co-founder of OthersideAI, that went viral. It's a very lengthy post, but it can be more or less summarized as follows:
We are in an uncomfortable period where a lot of people in the industry are sounding warnings, a lot of people in the industry are pushing bVery informative (and more than a little concerning), thanks for posting.
By now, I'm sure that a lot of people have seen the post from Matt Shumer, co-founder of OthersideAI, that went viral. It's a very lengthy post, but it can be more or less summarized as follows:
We are in an uncomfortable period where a lot of people in the industry are sounding warnings, a lot of people in the industry are pushing bthank you for sharing - i work with these models - particularly opus
it's been mind blowing what is done now after a few back and forths
it's not as good as he claims there's still constant problems - but an iterative back and forth with an ai will still get something that used to take a team of people a week to build can be done in an afternoon
I think I'm somewhat happy AI came about late in my life, rather than early. I'm really happy it wasn't around during my tech career. Not sure if I would have had much of a career.
The irony being that entry level engineers are being squeezed by AI but AI can't really do all the important hard stuff which takes expertise that entry level engineers will never get because they are being squeezed out by AI
I think we are gutting our future for what ultimately amounts to minor conveniences
Like, its very nice throwing a bunch of markdown docs at AI and having it spit out documentation but there is something to be said for performing the critical thinking necessary to aggregate and organize the information that conveys a deeper understanding about the decisions that went into the development of a tool or piece of software to the point that a normie can also understand the decisions that went into development
an iterative back and forth with an ai will still get something that used to take a team of people a week to build can be done in an afternoon
People seem to think there is magic in a certain kind of prompt but no, the magic is, you look at what it gave you and prompt it to improve. Or run it and get an error and feed it the error.
I am not swe I am systems engineer so I am never building software, I am doing automations and systems integrations and llms allow me to work with any language, any system, anything, ten times faster. It's completely revolutionary to my work and we don't actually know how to build good AI yet.
The irony being that entry level engineers are being squeezed by AI but AI can't really do all the important hard stuff which takes expertise that entry level engineers will never get because they are being squeezed out by AII think we are gutting our future for what ultimately amounts to minor conveniencesLike, its very nice throwing a bunch of markdown docs at AI and having i
Possible I suppose but I wouldn't bet a brass farthing on it being a serious problem.
It's more the case that now people can't deny that so many jobs are going, they are sort of moving to a position of denial that it will be more and more jobs going.
Soon enough to make worries about running out of the more expert engineers highly dubious.
a) AI is going to encroach more and more
b) It's a huge boost to the existing more expert engineers *
c) There are still some less expert people at companies coming through
d) It's enabling those not at companies. It's like the 80s on steroids for the kid in the basement or a small company.
* this is masked a bit by having to learn/develop the AI tools. That will pass quickly.
Pretty unlikely for a multitude of reasons
There is a reason twitter is inundated with "vibe coders" telling you how ****ed you are if you don't vibe code 16 hours a day yet there aren't any products being shipped
Its always been funny when people who have no clue what it takes to do things are trying to tell those deep in the **** what is or isn't
denial like the river runs deep. I dont doubt you will do well at the mocking though - even afterwards there is always great reason to mock.
Enjoy yourself ...
The irony of your statement being that you are not only ignorant, but arrogant
Maybe you could consider that I have to have these conversations regularly with experts on the subject
Please tell. Which AI experts do you regularly speak to and what do they have to say. Specifically on AI not encroaching very significantly into more expert areas over the next few decades
Handily if they are experts there should be some stuff you can quote
AI's wheelhouse is repetitive tasks, which is what a lot of the tools developed for cs do too. It just depends on what the demand for cse.
IBM reported last year that they let go 8k techs, replaced them with ai and now they've had to hire just as many back to deal with increased demand.
IBM reported last year that they let go 8k techs, replaced them with ai and now they've had to hire just as many back to deal with increased demand.
IBM did eliminate ~8,000 roles using AI (mostly HR/back-office).
IBM did end up hiring roughly the same number of employees back (in different departments) because AI-driven changes actually increased overall labor demand.
The rehired workers weren't necessarily the same individuals or the same jobs — the new roles were in higher‑skill, human‑centric domains.
These new hires were as follows;
More demand for customer-facing roles
More need for technical talent
Gaps where AI couldn’t replace human empathy, judgment, or complexity handling
Just another example of the changing job market, adapt or die..
Kinda like the old statement "computers will reduce the need for paper (circa 1980)", but instead it just enables humans to print more stuff.. lol
IBM is looking to be one of the winners from the boom in IT services. Even so it's not an overall head count rise, it's a repositioning from coding to customer facing, quality checking etc. In particular they need staff skilled in AI. Some of these will be very good jobs putting those who get them in a really good spot to have a lengthy career - fight hard to get one if you can. Don.t bet much on retiring at 50+.
It's good to see IBM doing well. Does anyone else remember the days when IBM didn't allow talk of competition and they boasted they never made anyone redundant. How about the old joke 'what do get if your merge IBM with Apple?"
Spoiler
IBM
live in seattle, nothing but massive layoffs every few months - the reason for this is always "repositioning those resources towards ai"
Dude, that's hilarious.
Spoiler
Not really. It's terrible.
I asked ChatGPT for IBM's revenue and profits over the last 10 years. Sure, they're playing great...
| **Year** | **Revenue** | **Net Income** | |
| -------- | ----------- | -------------- | ---------------- |
| **2015** | ~$81.7 B | ~$13.1 B | ([Wikipedia][1]) |
| **2016** | ~$79.9 B | ~$11.8 B | ([Wikipedia][1]) |
| **2017** | ~$79.1 B | ~$5.7 B | ([Wikipedia][1]) |
| **2018** | ~$79.5 B | ~$8.7 B | ([Wikipedia][1]) |
| **2019** | ~$77.1 B | ~$9.4 B | ([Wikipedia][1]) |
| **2020** | ~$73.6 B | ~$5.5 B | ([Wikipedia][1]) |
| **2021** | ~$57.3 B | ~$5.7 B | ([Wikipedia][1]) |
| **2022** | ~$60.5 B | ~$1.6 B | ([Wikipedia][1]) |
| **2023** | ~$61.8 B | ~$7.5 B | ([Wikipedia][1]) |
| **2024** | ~$62.8 B | ~$6.0 B | ([Wikipedia][1]) |
