David Simon, creator of ‘The Wire’, being interviewed by Ari Shapiro (NPR)
SHAPIRO: Ok, so you’ve spent your career television without AI, and I could imagine today you thinking, boy, I wish I had that tool to solve those thorny problems…
SIMON: What?
SHAPIRO: … Or saying…
SIMON: You imagine that?
SHAPIRO: … Boy, if that had existed, it would have screwed me over.
SIMON: I don’t think AI can challenge what writers do at a fundamentally creative level.
SHAPIRO: But if you’re trying to transition from scene five to scene six, and you’re stuck with that transition, you could imagine plugging that portion of the script into an AI and say, give me 10 ideas for how to transition this.
SIMON: I’d rather put a gun in my mouth.
I’ve had this exact conversation in real life.
I do IT work. Someone who was a heavy AI user asked me ‘how was I integrating AI into my workflow’ as if it was a foregone conclusion that I AM integrating AI. I told him I wasn’t, that AI may be useful for finding answers to some things or photoshopping funny pictures but most of what I do is intellectual, specific to certain situations and requirements, which AI generally doesn’t understand. That I can’t risk AI making actual changes to my environments, so if I let AI write a plan and then I proof it myself and then do it, it’s easier just to do the work myself. So that leaves the only thing AI could do for me is writing, and I’m good at writing, so I overall have little or no use for AI.
Dude looked at me like I was an alien. Like he legitimately could not understand why an intelligent person would voluntarily reject AI. Like I said I rejected cars and I ride a horse to work. Does not compute. Here’s this supposedly smart person who knows more about tech than the next 2-3 people in the office combined, and this person is completely rejecting the holy grail.
I think everyone should think of AI as that you have random access to the knowledge of reddit. Sometimes it’s good, sometimes it’s a troll, and sometimes no one will answer. You also have to have knowledge to understand what it’s saying most of the time.
at least reddit doesn’t (well, didn’t) consume a nonillion jiggawatts just to produce a half-shitty answer
But, you have to specify it isn’t actually redditors seeing the question and posting an answer. AI doesn’t actually understand questions.
It’s more that it sees your question and generates text near that question that wouldn’t look out of place on Reddit. Sometimes what it generates is good, sometimes it’s what a troll would say, the key thing is that it’s believable but not accurate.
Right, I think we’re agreeing. That is exactly like reddit.
It’s tough not to form a superiority complex these days when you hear of people who can’t compose emails by themselves.
Yeah, and I think it’s a symptom of a cultural problem where people think prosperity means essentially not having to do anything. The ideal is to sit on your ass and consume, while never having to create.
But when you don’t do anything for yourself and have AI or hired help do it so that you can go back to scrolling the dopamine treadmill, you eventually don’t know how to do it even if you wanted to. Losing knowledge, skills, and attention span means getting more stupid, kinda by definition.
This negative feedback loop is a form of conditioning, and it is not in our best interests. It is in the interests of advertisers, psychological manipulators, tech barons, and others who are generally the worst.
In my case, the more I have consciously decided to take care of my own shit – in various aspects of life – the better I have felt. It’s not the one secret to a fulfilling life, but it’s an important one. And yes damn it, I sometimes use dashes.
My boss had three of us on a Teams call for nearly an hour and a half yesterday because he needed our help to make one Power Point slide.
To be fair, power point - if you don’t regularly use it and especially if you’re using a ‘template’ someone made - is a miserable experience.
Add four words to a sentence - why did the fucking font size change and why did an unrelated text box shrink?!?!? I just need to add ‘this has been approved’ to my status deck.
Why can’t I update a blob of text? Oh, it’s in the Slide Master, which has exactly one instance of this slide type in the deck.
/rant. I haven’t lowered myself to using copilot to make these changes, though.
Meanwhile I just plugged Claude into a production cluster to perform analysis and I accepted its suggestions for an immediate fix, while it also made those changes to my Ansible setup and committed them to git.
I think AI is great for some things, but imo it should NOT be wasted generating funny pictures or videos.
I think you’re one un-caught hallucination away from having claude eat the production database. I get that you think you’re approving suggestions, but there are many examples already of “<llm> ignored my directive of don’t make changes and then said sorry” out there.
If a prompt can make changes, then at some point a prompt will make bad changes. It hasn’t happened to you yet, apparently, but you’re playing Russian Roulette with the business.
I do agree it has some use cases, but shouldn’t be used to make pictures, video or writing text a human claims as something they produced.
We’ll see, I like to live life dangerously, especially when my org is pushing AI down our throats
Same. Planning-with-files has saved my ass more times than I can count.
You’d think that last sentence would make people understand HOW you got that knowledge…