|
Ed Zitron@zitron.bsky.social |
Are you KIDDING ME? This is exactly what they've been selling! The whole thing is that AI is meant to magically change your business!
43 replies 56 reposts 329 likes
|
Kirk Murphy
@kirkmurphy.bsky.social
[ View ] |
They’re trying to imagine having a friend.
0 replies 0 reposts 0 likes
|
MurkyMurk
@naught2shabby.bsky.social
[ View ] |
The thing every B-schooler "learns" besides their Myers-Briggs result: Every new tech is a silver bullet that will improve profits if shouted loudly enough and injected everywhere. *Shiny* Tech is Emperor's Clothes for every Exec seeking relevance or suffering from profit-related FOMO.
0 replies 0 reposts 0 likes
|
PeoriaBummer
@peoriabummer.bsky.social
[ View ] |
The business model of getting middle management to pay for enterprise accounts that end up on the budget for decades without providing value.
0 replies 0 reposts 1 likes
|
Tom VanAntwerp 🤷♂️
@tomvanantwerp.com
[ View ] |
Speaking as a former IT guy, Richard Seroter sounds like a sysadmin who is so fucking tired of people not organizing their goddamn files and is happy to find any excuse for people to stop storing important documents in the Downloads folder or, god forbid, the Recycle Bin.
0 replies 0 reposts 1 likes
|
J Frederick
@jfrederick.bsky.social
[ View ] |
If AI could actually get a lot of big companies to fix their source system data issues (and their downstream/integration chaos), then it would be worth it for that ALONE. But fixing that is hard and expensive and you’ve just spent all your money on magic beans.
0 replies 0 reposts 0 likes
|
Hans, Faffing
@hansfaffing.bsky.social
[ View ] |
ahahahaha
0 replies 0 reposts 2 likes
|
art toast
@toasty.cx
[ View ] |
i legitimately misread that as Richard Scroter and honestly completely lost my ability to comprehend anything else
0 replies 0 reposts 7 likes
|
The Magenta Egg
@emptymind.bsky.social
[ View ] |
Yeah - business “data” isn’t ever all set up in a tidy spreadsheet for machine consumption. It’s always some compost pile of text, spreadsheets, pdfs, voicemails, images, emails, messages, socials, whiteboards, and ideas in people’s heads all competing against each other. Good luck AI’ng that.
0 replies 0 reposts 0 likes
|
Jonno
@jonnosan.bsky.social
[ View ] |
Another way of looking at this is, a whole bunch of techos have found a way to get important boring data cleanup work funded, by telling their bosses its an AI project
2 replies 1 reposts 8 likes
|
P
@intothejunction.bsky.social
[ View ] |
I would like to be a Chief Evangelist. Give me seven million dollars to go on TV and tell everyone that all this broken bullshit will definitely be really good at some undetermined future point.
0 replies 0 reposts 1 likes
|
Klern Shibdall
@hankdacat.bsky.social
[ View ] |
Also, it doesn't work that well even if your data is super organized. It just writes wordier emails and tells you how to find stuff that is already easy to find (bc the data is well organized!!)
0 replies 0 reposts 1 likes
|
Ben Peake
@reddishraven.bsky.social
[ View ] |
This really feels Like Lt. Gorman in Aliens. Ripley : How many drops is this for you, Lieutenant? Lieutenant Gorman : Thirty eight... simulated. Private Vasquez : How many *combat* drops? Lieutenant Gorman : Uh, two. Including this one. The real world is a xenomorph waiting to eat you.
0 replies 1 reposts 7 likes
|
Insipid Twaddle
@insipidtwaddle.bsky.social
[ View ] |
"We have developed the ultimate business product: an artificial toddler. Without constant supervision it will ruin anything valuable comes in contact with, it steals constantly and without conscience, and it struggles to differentiate between reality and its imagination. Please give us billions".
1 replies 4 reposts 12 likes
|
Matty
@matt78.bsky.social
[ View ] |
the one thing “AI” has been useful for, for me, has been to write a dry one to two sentence summary of a data table that I then critique. But the data needs to be extremely well structured. Even then it needs a very specific prompt, or it may discuss results from 2019 rather than the most recent.
0 replies 0 reposts 0 likes
|
Frisky Dingus 🍉
@friskydingus.bsky.social
[ View ] |
I will say all this AI failure to deliver has been great for me, winning disagreements with my partner who thinks tech is going to lead to a utopia somehow.
0 replies 0 reposts 5 likes
|
Dr. Benway
@docbenway.bsky.social
[ View ] |
What does “six units of AI” even mean?
0 replies 0 reposts 0 likes
|
Annikai
@annikai.bsky.social
[ View ] |
Lol I love it being contingent on databases businesses having correct data, something which we all know businesses are great at and not at all regularly out of date. Yep, I've never spent an hour long meeting troubleshooting something only to find the issue was bad data
0 replies 0 reposts 0 likes
|
Cormacaroni
@cormacaroni.bsky.social
[ View ] |
Garbage in, garbage out still undefeated
0 replies 0 reposts 0 likes
|
Stephanie goes by Stephanie
@pagnificent.bsky.social
[ View ] |
Are you kidding me?! You couldn’t copy and paste to provide alt text?!
0 replies 0 reposts 2 likes
|
Populares Stan Account
@mgraalum.bsky.social
[ View ] |
Most clients D365 applications are held together with duct tape, getting the house in order isn't even plausible
0 replies 1 reposts 12 likes
|
Derek of the Edmund Fitzgerald
@giromide.online
[ View ] |
any AI built off my employer's internal wiki would regularly offer anachronistic advice on a variety of critical applications because stale pages aren't archived for fear of never being able to retrieve (we sacked the admin)
0 replies 0 reposts 0 likes
|
Mya Riemer
@myariemer.bsky.social
[ View ] |
I have yet to encounter a corporate enterprise that truly has its data in order, and they also do not find data management at a sufficient level to improve that. DM is hard work — just determining where data originated is often impossible. Just throw AI in the mix, though, it’ll solve everything.
1 replies 1 reposts 1 likes
|
Jason
@jasonh.bsky.social
[ View ] |
Was at a conference recently. AI was a big topic and a major through-line was data prep. "Majority of our deployment was data prep." Unless you have perfect data sets, perfect word docs w tables of contents and labeling, PDFs properly converted from source, these "tools" are functionally worthless.
0 replies 0 reposts 2 likes
|
miakepia 🇵🇸
@miakepia.bsky.social
[ View ] |
now THIS is comedy
0 replies 0 reposts 0 likes
|
RTOIT
@rtoit.bsky.social
[ View ] |
I keep thinking about Pharrell Williams' "Cash In Cash Out" when I read these stories. Trash In, Trash Out. (Sorry not sorry about dropping that ear worm, and you're welcome)
0 replies 0 reposts 0 likes
|
|
@srmadscience.bsky.social
[ View ] |
I’ve spent my career working with data. It’s not possible to clean all the data in a large corporation. It’s messy because reality is messy. So other than infinite electricity and perfect data what else do I need to do “AI”? A shelf for storing snake oil?
0 replies 0 reposts 2 likes
|
MindCrime
@mindcrime.bsky.social
[ View ] |
Is ai good at any of the real difficult computing problems like the Traveling Salesman Problem?
4 replies 0 reposts 0 likes
|
Machinedad ↙️↙️↙️
@machinedad.bsky.social
[ View ] |
The most frustrating thing is that every single one of my friends and colleagues who have applied ML to real world business problems, including junior engineers who have spent maybe a year on the job, could have and did tell people this.
1 replies 0 reposts 2 likes
|
putagonia
@newreply.bsky.social
[ View ] |
google needs companies to neatly categorize all its data and hand it over to them. they absolutely have no ball in the game here except for helping out dear customer ofcourse
0 replies 0 reposts 1 likes
|
|
@heim2yankel.bsky.social
[ View ] |
In the script it says "Hilarity ensues".
0 replies 0 reposts 1 likes
|
Alan Bleiweiss
@alanbleiweiss.bsky.social
[ View ] |
"magic"
1 replies 0 reposts 3 likes
|
Jim Kosmicki
@jkosmicki.bsky.social
[ View ] |
It's not OUR product that is shit, it's YOUR bad data
0 replies 0 reposts 0 likes
|
®PG
@realphilgold.bsky.social
[ View ] |
Boiling the ocean
0 replies 0 reposts 0 likes
|
Malaclypse the Fetus
@omfproductions.bsky.social
[ View ] |
Kind of an open secret that "good data" has been at best a tertiary concern in the C suite for at least 20? 30? years now?
0 replies 0 reposts 1 likes
|
Colon Musk, CEO of GaceX
@themilfmag.net
[ View ] |
hes the most downest bad i ever seen someone be for friends
0 replies 0 reposts 1 likes
|
Julian
@jl8e.bsky.social
[ View ] |
Sounds like they’re blaming the customers for not getting anything useful AI cannot fail; it can only be failed
0 replies 0 reposts 3 likes
|
Drifting
@softapocalypse.bsky.social
[ View ] |
Somehow I feel very reminded of this old tale: en.wikipedia.org/wiki/Stone_S...
1 replies 0 reposts 4 likes
|
Ian Rennie
@theangelremiel.bsky.social
[ View ] |
I wonder if companies like OpenAI realise how much long term damage they have done to their brand by lying like fuck about what AI is good for.
3 replies 2 reposts 25 likes
|
Ultraklystron 🏖️ Karl Olson
@karlrolson.com
[ View ] |
It has been delightful seeing people & companies realize they've been rope-a-doped into buying a per seat license for RNG auto-complete that typically only works if they've already been writing top shelf code from top to bottom.
0 replies 0 reposts 10 likes
|
Julia S.
@booktweeting.bsky.social
[ View ] |
literally any actual machine learning researcher could have told them* this if they had bothered to ask *captains of industry** **also credulous media
0 replies 0 reposts 4 likes
|
bangaroo
@bangaroo.io
[ View ] |
wait… wait……. if I didn’t know any better I’d think this mean they’ve consistently overstated the impact of spicy autocorrect and are now trying to walk that back to reality…
0 replies 0 reposts 1 likes
|
Robert Brydon
@robertbrydon.bsky.social
[ View ] |
Seems pretty reasonable. How could you outperform an AI system, assuming you first normalize and sanitize all of your business data, create reports with robust mathematical regression models, and put those in a well indexed location? Other than doing all that and skipping the "AI", of course.
1 replies 1 reposts 15 likes
|
Double Waingro
@palinode.bsky.social
[ View ] |
It reminds me of the scene in Go where Sarah Polley sells fake tabs of acid and says “It helps if you smoke a lot of pot with it - like, a lot”.
0 replies 0 reposts 4 likes