"Some creative jobs maybe will go away," Murati said in an interview at The Thayer School of Engineering at Dartmouth University last month, as quoted by Fortune. "But maybe they shouldn't have been there in the first place."
Why should the jobs have not been there in the first place? If someone is willing to pay another person to perform a job, so be it. It's their money. To say it maybe shouldn't have existed in the first place seems.. narrow minded.
We used to have a job called computer, and people were employed as computers in living memory. When we say that a job does not exist, we speak about how we wish to walk going forward.
OK but you could say that about practically any job - given enough time and technical progress. But in this context we are talking about these few jobs where people really enjoy what they are doing and actually create something genuinely new. If all artist disappear and what we're left with is GenAI feeding on GenAI, it's a bit sad because art is an important part of our culture.[0]
[0] Well, Murati could say, "Maybe it shouldn't be in the first place"...
There is a innate human affinity for art. The question is whether young artists will be taking on the prestigious internship of drawing Meg Griffin's eyes. If people love art then let them draw.
Watch videos with OpenAI execs without sound. It’s quite obvious they’re making up their outlandish claims as they go along. It makes me kind of worried actually that there’s not much more gas in their particular tank and the whole “transformers can do everything” thing is reaching a plateau.
OpenAI is not the only game in town. And Anthropic seems to think they've got another trick up their sleeve since their current leading model is only their mid-product line with plans to release the top tier Opus later this year.
LLMs are nothing more than statistical token prediction engines, feed it bullcrap and it will statistically provide bullcrap. There is no novel thought behind them, and they're marginally more complicated than your phone's keyboard Next Word Predictor.
So, yeah, they'd fit in place of an McKinsey alumni sitting in a C-Suite transparently.
This might seem like a childish retort, but I genuinely think that the job of CEO should definitely be given to AIs. I mean, it's both the most expensive and most risky job of them all, and there's nothing a CEO does that even the current LLMs couldn't handle.
I suspect an LLM could possibly handle generating the text in a quarterly report, as well as many mundane tasks, but not the question and answer session on the earnings conference call because nuance and judgment is required to communicate critical information to the market.
Besides the non-trivial task of communication, I have the impression that there is a general belief that companies with a bad CEO tend to die and companies with a good CEO often thrive.
How would you make sure that the LLM is a good CEO as opposed to a CEO that steers the company toward failure?
This is an idiotic take for 2024. A CEO has to hold an enormous amount of information in-brain and make decisions based on all of it, and handle interpersonal relationships, and manage people, etc. LLM’s can barely handle the information in the context windows they have now.
Why should the jobs have not been there in the first place? If someone is willing to pay another person to perform a job, so be it. It's their money. To say it maybe shouldn't have existed in the first place seems.. narrow minded.
reply