Every review of AI is like: This isn’t very useful and it makes it harder to do simple tasks than if I just did it myself. Riddled with errors. Can’t see anyone using this. I feel stupider having tried this. That said I believe it will soon make human cognition redundant and eradicate us all.
Comments
"It doesn't actually perform that great, you really can't do truck things with it, and I sliced my hand on the trunk, but it's a revolutionary vehicle that will absolutely-"
(Right before a car wash bricks it, or a tow truck shears off an axle, or...)
It vastly increases my capacity to get research done as a result.
But to the average person it's only useful because the big tech companies made search engines shit.
btw...i am so glad you are here. keep asking your sports analyst friends to join you here♥️.
When we finally got it we realized it was dumb AF
It’s changed my life so that I now have to duck whenever I go into the living room so my son doesn’t smack me in the face while he’s playing gorilla tag
Get onboard before you get left behind!!
Use as a training data set Howard Zinn‘s history of the United States.
Brother. No.
It was nice to gather the feedback, have AI summarize into categories & write up suggestions on how to action each category. Probably 30 mins, at least
But as long as we're clear: it's not actually artificial intelligence. It's an analysis program. There's no actual "intelligence" involved.
It’s 2025: Where is my Teddy?
1. It is a tool to be used to assist and augment humans in their work and in some specific instances, automate work.
2. To your point about the errors and my point above, you should never turn a process, even an automation one, entirely to AI. There's a concept called Human In The Loop (HITL) and it states that any AI model/process should at some point in it workflow, have
People who think otherwise don't understand this technology, which would point to, a lack of curiosity.
The main product of Silicon Valley is hype-fueled stock options.
Antibodies by Charles Stross
https://www.baen.com/Chapters/9781625791870/9781625791870___2.htm
I’m not necessarily saying that’s a good thing, but judging AI tools based on their performance today, and deciding that AI is all crap (which you didn’t do, but many are), is probably premature.
I’m afraid you’re right. It will finish
the job the internet started.
It’ll do it precisely because it’s too dumb to handle complex systems.
https://what-if.xkcd.com/5/
Also like, using it to write *anything* means spending your time making sure it didn't fuck up
You can't treat "generative AI" as a synechdote for the wider set of machine learning. It's a cul de sac, astonishingly worse than GIGO - quality goes in, quality-shaped garbage comes out, misleading until examined.
Scientists know not to just trust what the AI suggests. They test and often improve the molecules before exposing humans to the output.
Tech bros just barf extremely dumb LLMs all over our lives.
Now, I am just an anonymous schmo
and here https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3542
There are definitely useful applications of AI/ML methods in biology, but there is a lot of hype, too.
https://www.sciencedirect.com/science/article/pii/S135964462400134X
You are describing iterative AI. Hugely helpful at quickly exploring design spaces to find improvements.
Iterative AI is very different from the current fad of generative AI, which is largely slop.
RegEx is especially easy to be screwed up by generative AI.
There's no regex or SQL that AI could do better than a proper Google search and a manual.
Okay, but by a generative LLM? Or via iterative machine learning?
Same with code, imagine just generating something, making sure it compiles and assuming it's correct because it compiled.
That's what they're telling consumers AI can do.
Why do LLMs expect publishers to?
come back when it doesn't boil the oceans doing so
bit like saying some of the people on cruise ships deserve a holiday, so ignore the pollution otherwise
https://www.jackhorne.net/some-useful-a-i/
So it's not usable, a total idiot, limited by our idiocy to grow any further unless we do, and the biggest threat we face?
Are the critiques written by it? 😂
AI still has a very long way to go.
I say this as a Software Architect with 25yrs experience.
It cannot, will not be AI without senses--hearing and vision at the bare minimum. Just Google, but even worse
Is the juice worth the true cost of the squeeze? Maybe. Trillion-dollar industry? No.
In my case, I'm likely being auto-rejected because I'm trying to relocate so my application has a city a couple hours away
But that's not generative AI. That's a just a basic filter that is unfortunately filtering me out.
The general public is already primed to believe in a certain type of “AI” that’s far more useful to the grift.
...and then he spits out the longest list of custom commands and edits you've ever seen
e.g. computers when they started were not necessarily more efficient than a lot of paper record keeping. peeps who got into them were mocked, people didn't always see the potential.
now we have computers in our pockets everywhere we go
https://m.youtube.com/watch?v=EUrOxh_0leE&t=3245s
Maddening.
*There will be a LOT of ppl who won’t be able to tell the difference and that is the scary part imo.
This is so great to see as your response.
It really helps to illustrate your complete lack of self awareness & your inability to discern “mockery” from “anger”.
I appreciate being reminded on a regular basis that the entirety of online crypto bros/AI humpers share a single brain cell.😂
Soon we won’t need to learn anything! We can just copy and paste it!
But right now, generative AI when used for things like script writing is dull and predictable.
It doesn’t have to be all or nothing
So useful that corporations will USE it to replace as many humans as is physically possible.
Used properly and it can be a fantastic tool. Used improperly and it could quite easily kill us all.
What happens when Elon plugs AI into the DOD top secret databases? Weapons?
Ouch. 😬
The only thing it’s actually good at is switching between vernaculars.
Fucking excellent.
-the AI story.
Because it has no regard for the truth.
Gemini: I'm sorry, I can't do that.
Me: Gemini, how much battery do my ear buds have left?
Gemini: I'm sorry, I can't do that.
W H A T ? ! ? !