I don’t know but implication the other poster is making is “a human can write 2 articles, a Ai can write 5, I’m being asked for 5 which is impossible. I can use Ai and risk trusting it or not meet my required outputs and also get fired.”
I made up those numbers but that’s the accusation. You are damned if you use the Ai to meet your goals. You are damned if you don’t meet your goals.
Possible but back when smoking was much more common (not like 2010s, I mean like 1930s when a pack was included in mre kits given to soldiers in ww2 and the Flintstones cartoon had sponsored segments where Fred and barney had “smoke breaks” and talked about the “smooth and delicious taste” of I think it was camel cigarettes ) There were many more companies making cigarettes. It was only really consolidated in I think the 80s due to lawsuits for causing cancer, false advertising, lying to congress etc. Most of the companies couldn’t afford all the fines and were bought out.
Really? If 1% of customers of a restaurant got food poisoning… Well that’s exactly why we have a health department and food inspectors
Now if by slop you just mean “not delicious” people would probably stop going and stop recommending. Some places may be able eat that loss(pun intended) but many would go out of business. Related, sorta, there have studies that one aspect of McDonald’s that added to its huge success is consistently being same basically everywhere. Yeah there’s some regional differences but you can eat a big Mac basically everywhere. And they are basically the same.
No, I don’t think it was an opinion. My point is in a lot of cases something really bad would never hit 1%. Like commercial airlines, most do several hundreds of flights a day(just a guess but for the large ones seems reasonable across all airports). If one airline had say 3 crashes in a year I can’t imagine them still being active.
Population of us in 2024 was 340,000,000. If it effected 1% that would be 3.4 million people. That’s like 1/2 of the holocaust victims.
I believe most companies would be done at 0.01% if that. Only company I can think of with that body count is nestle… I don’t even think Raytheon has killed that many.
I think you drastically overestimating 99% vs 1% at scale. For example look at the Chicago Tylenol case. Atleast 7 people died. 31 million packages of Tylenol were recalled, the entire line of powder filled capsules was eliminated, nationwide redesign of medication packages was done. https://en.wikipedia.org/wiki/Chicago_Tylenol_murders
Rubber duck programming
I want to read your 14 terabyte book about unicorns swimming with dolphins… I don’t think it’s text
I’m sorry, I think we agree and I explained it wrong.
I was mostly joking in my response.
That said you are mostly/kinda right. If person A hits a pedestrian, person b, then person b can sue person A’s insurance. Two major issues… The insurance can afford much better lawyers than person B and can afford to drag it out longer, resulting in person B not getting the compensation they should. Also person A has/will pay more in monthy charges than the insurance will likely ever pay out. That is why they make profit.
It’s a loss for both parties to have insurance.
Yeah but if you are injured or killed in a warehouse the warehouse owner may be sued or held liable which hurts rich people. If you get killed on the street only the poor suffer.
/s
“don’t copy that floppy”
I built a hundred bridges but am I known as John the bridge builder? NO! I built a dozen a castles but am I known as John the castle builder? NO! But I fuck one sheep…
True, nothing is perfect.
If I need a life saving operation and given the choice of a Harvard graduate at random or a random person in time Square… I know who I would choose
OK the person from time square but that’s probably just my depression speaking.
This is a case where I wish we didn’t have a “jury of peers”. Most people don’t even know dompemine is a thing or it’s effects. Most people don’t know the effects a disfunctional family can cause.
No, I don’t understand them completely either. There are experts in these fields who have done valid research and run well documented studies to better understand these topics. They are who should determine this. No YouTube videos or a blog post don’t count. I mean actual real studies that stand up to scientific rigor.
I feel that YouTube/facebook played a part. But I want that to be true due to my inherit bias. I have no data or numbers to back that up. Experts probably do. There have likely been done studies about these things.
I know know there is a thing called a Skinner box but that was over 50 years ago and I’m sure we know more. That said I’d bet the average person on the street doesn’t even know that and shouldn’t be determining this case.
Ask your McDonald’s worker for something to release some dopamine after a long work day, based on their answer do you want them setting precedent on this topic?
Look, I know I am no longer young and hip but calling me a dinosaur hurts my feelings…
I’ve said it in other similar articles, it depends on your definition of “Ai”. Marketing is insisting on “powered by Ai” on all sorts of products but that doesn’t mean llm. Some “Ai” may be useful, like the old on-star system on some cars. They had multiple inputs to determine if a crash occurred and alert emergency services if it did… Could be classified as Ai.
Your toothbrush example could be useful, if just finished cleaning cycle and battery level below 20% and not on charger, emit a beep to alert user to charge tooth brush. More advanced, and I doubt these are the case but would be cool and useful. If the could detect blood during a cleaning cycle, alert user to contact a dentist for possible gum disease. Or detection of a new/growing hole in a tooth based on the defection of the bristles and alert user to see a dentist for possible cavity.
Without a solid definition of “Ai” this is all marketing talk. Basically all “Ai” takes multiple inputs and then generates an output based on those inputs. If you say it must generate it based on a llm, then what about the image generating ai? If you say it must have natural language as in an input most of these “Ai” products don’t qualify unless they expect you to say “ok, sonic care end cleaning cycle” with a mouth full of toothpaste and water. Technically an argument could be made that a check engine light on a car is “powered by Ai”.
The only definition which goes against this and I would agree to is a system where given an identical set of inputs the output is not always(or ever) the same. If that’s the case then I am going to start making “Ai” powered lava lamps…
Also true
Think back… XP had many complaints at release too, wasn’t till sp2 or sp3 it was actually decent.
Microsoft can never name products… Xbox, Xbox 360, Xbox one, Xbox series x, Xbox series s… Hell I’m currently watching agdq and can’t tell you the best one, imagine a grandmother trying to buy a gift for her grandchild. Atleast it’s Playstation 1, Playstation 2, Playstation 3, Playstation 4, Playstation 5 obvious which one is newest. Or Windows 3.1, nt, 3.1 with workgroups, 95, 98, 2000, me, xp, 7, vista, 8, 10, 11…wtf?
Ouch! Tell her I’m sorry, and I’m sorry for you too. All the accountants I worked with did alot more than just reports. Not to mention that sounds great until the Ai says 2+4 =2*4 and now the company owes 20 billion on taxes…
Plus in a lot of cases people don’t submit records in identical format, the number of excel workbooks I’ve seen where the data was on “sheet 2” for some unknown reason…
Maybe its just me, I always provided raw data on sheet 1, analyzed data on sheet 2, and if needed complicated formulas on sheet 3. I would be willing to bet their Ai would break on that format.