Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Slack AI

slack.com

Post reply on HN
Print view View original

11–20 of 54 posts

Re: Slack AI

#11

Post by ivanjermakov »

morkalork wrote: post #5

Man, it has got to suck being a fresh out of school CS grad. Not only is the industry fucked, if you get a job, you may well be fully remote and WFH which already kills social connections in the crib. Now add slack AI to the mix and you've got the Black Mirror-esque nightmare of: Has your manager really read your message? Have they ever read any? Or is everything you say just being shouted into an abyss, with only an…

Isn't it responsibility of a manager to treat employees well? Careless manager wouldn't read their messages regardless LLM or not.

Re: Slack AI

#14

Post by SystemOut »

pfych wrote: post #7

I wonder how long it will be until we hear about a company accidentally doing something because the Slack AI summary / AI search incorrectly paraphrased a request. I really wish AI companies would call hallucinations what they really are - Mistakes. Using wishy-washy words like that make non-techy people trust outputs of LLMs much much more than they should. "Oh it just hallucinated - how silly!" instead of "It got i…

I think it's only going to get worse as the LLMs get just that much better that people stop thinking more critically about what they generate. Right now I think many people still verify or read through the results to make sure they sound right. What happens when people get lulled into complacency and then it really goes off on a bender.

Re: Slack AI

#15

Post by beardedwizard »

mistrial9 wrote: post #9
phildenhoff wrote: post #4

Is this the beginning of the end of Slack? Unrelated: I forgot Slack was owned by Salesforce, who also own Heroku. Why am I still using Slack?

.. because it is a mass illusion that you have individual choices in the matter. Companies and some govt agencies are required to have a monitored, logged and legally responsive chat provider. You are not asked if you want to use Slack, you are required to use Slack. hint- expect much more of that

Expect much more of it from who? What are you suggesting changed? Among all the spying enterprise apps, why slack? Afaik, engineers brought slack into many companies and pushed back on alternatives like teams.

Re: Slack AI

#16

Post by LeoPanthera »

pfych wrote: post #7

I wonder how long it will be until we hear about a company accidentally doing something because the Slack AI summary / AI search incorrectly paraphrased a request. I really wish AI companies would call hallucinations what they really are - Mistakes. Using wishy-washy words like that make non-techy people trust outputs of LLMs much much more than they should. "Oh it just hallucinated - how silly!" instead of "It got i…

> I really wish AI companies would call hallucinations what they really are - Mistakes.

At the bottom of every ChatGPT conversation: "ChatGPT can make mistakes. Check important info."

Re: Slack AI

#17

Post by duskwuff »

pfych wrote: post #7

I wonder how long it will be until we hear about a company accidentally doing something because the Slack AI summary / AI search incorrectly paraphrased a request. I really wish AI companies would call hallucinations what they really are - Mistakes. Using wishy-washy words like that make non-techy people trust outputs of LLMs much much more than they should. "Oh it just hallucinated - how silly!" instead of "It got i…

> because the Slack AI summary / AI search incorrectly paraphrased a request.

Or because it misinterpreted a question (or a joke!) as a factual statement.

Re: Slack AI

#18

Post by threeseed »

sciolist wrote: post #6

Even if AI tools (e.g. search or summarization) aren't perfect, neither are our current manual methods. Right now I feel there's a strong negative sentiment towards mistakes made by LLMs vs humans - when/if that sentiment fades, it'll be interesting to see how much work can be delegated to LLMs.

The difference is in how you make mistakes.

Humans have the ability to explain the thought process that led to the mistake, can provide confidence intervals, can learn from the mistake and can maintain a coherent position over time.

And right now the only work being delegate to LLMs are those where mistakes are tolerable e.g. copywriting, customer service.

Re: Slack AI

#19

Post by sadops »

pfych wrote: post #7

I wonder how long it will be until we hear about a company accidentally doing something because the Slack AI summary / AI search incorrectly paraphrased a request. I really wish AI companies would call hallucinations what they really are - Mistakes. Using wishy-washy words like that make non-techy people trust outputs of LLMs much much more than they should. "Oh it just hallucinated - how silly!" instead of "It got i…

I really wish we'd stop anthropomorphizing these things. They're not hallucinating, the program simply returned an error. Remember, y'all, this stuff is just Markov chains with extra steps. An incorrect answer from any other computer program is called an "error", and apparently LLMs are the only piece of software in which significant amounts of them are tolerated, because they look just like a mechanical Turk, and are sufficiently convincing as to be a really neat parlor trick. But write a program that generates random integers and returns -1 for even numbers, and it's the exact same thing. LLMs are crap, because programs aren't supposed to return errors.

"On two occasions I have been asked, — 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?'... I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.

Re: Slack AI

#20

Post by sadops »

LeoPanthera wrote: post #16
pfych wrote: post #7

I wonder how long it will be until we hear about a company accidentally doing something because the Slack AI summary / AI search incorrectly paraphrased a request. I really wish AI companies would call hallucinations what they really are - Mistakes. Using wishy-washy words like that make non-techy people trust outputs of LLMs much much more than they should. "Oh it just hallucinated - how silly!" instead of "It got i…

> I really wish AI companies would call hallucinations what they really are - Mistakes. At the bottom of every ChatGPT conversation: "ChatGPT can make mistakes. Check important info."

[flagged]
Post reply on HN