Ancestors

Toot

Written by May Keable 🏳️‍🌈🏳️‍⚧️ on 2024-10-27 at 15:12

I think that fact that we're using AI to write emails because we find it hard and AI to summarise emails because we can't be bothered to read them suggests that we should take a look at how we communicate rather than boiling the oceans to have LLMs hallucinate at each other on our behalf.

=> More informations about this toot | More toots from Keab42@kind.social

Descendants

Written by Mx Verda on 2024-10-27 at 16:25

@Keab42

As an avid “please only communicate with me in text — oh wait nvm you prefer calls and there is no other option? Ok, I’ll just have a nauseous dissociative spell after your unannounced auditory terror into my life. Nbd” person,

Yes.

Please. Please get the NTs the learn to read.

=> More informations about this toot | More toots from MxVerda@lgbtqia.space

Written by Arnan on 2024-10-27 at 16:30

@Keab42 No, people just need to stop being lazy and make an effort.

=> More informations about this toot | More toots from arnan@mas.to

Written by Sally Strange on 2024-10-28 at 23:11

@arnan @Keab42 laziness is the mother of invention

=> More informations about this toot | More toots from SallyStrange@eldritch.cafe

Written by Jo being generally creative. on 2024-10-27 at 17:43

@Keab42 clearly so much is useless and time wasting filler. Most autistics already recognize it as blather.

=> More informations about this toot | More toots from JoBlakely@mastodon.social

Written by :jan:‍:abreath:🌬️:dandelion: on 2024-10-27 at 18:28

@Keab42 I'm hiding this in my email signature at work

=> More informations about this toot | More toots from Crazypedia@pagan.plus

Written by Tom Walker on 2024-10-27 at 19:57

@Keab42 The solution is to only ever write one-line emails

=> More informations about this toot | More toots from tomw@mastodon.social

Written by My name is Gordo (not really) on 2024-10-27 at 20:13

@tomw @Keab42 Considering that number of people that only read the first line, this is pretty sound logic.

=> More informations about this toot | More toots from gordoooo_z@hachyderm.io

Written by Stacey Cornelius 🇨🇦 on 2024-10-27 at 20:13

@Keab42 One thousand percent this.

=> More informations about this toot | More toots from StaceyCornelius@zeroes.ca

Written by Pier344 on 2024-10-27 at 20:38

@Keab42 having AI do our office work is like having machines replace us in agriculture, farming, mining, manufacturing and transportation. It means we are not made for capitalism, but rather for enjoying nature and freedom while machines work for us. 24 hour workweek should be the norm (4 days).

=> More informations about this toot | More toots from pier344@mastodon.social

Written by EndlessMason on 2024-10-28 at 20:34

@pier344

Nah. Those other machines do useful stuff - corn or potatos or pants or overnight flashlights - these are useful

A status email from a weekly meeting that people use as a catch up / down time and read minutes that are just "everybody still working on the thing from last week", not so much

@Keab42

=> More informations about this toot | More toots from EndlessMason@hachyderm.io

Written by Pier344 on 2024-10-28 at 21:12

@EndlessMason @Keab42 agreed and followed, thanks for that contribution, appreciated

=> More informations about this toot | More toots from pier344@mastodon.social

Written by Megan on 2024-10-27 at 20:41

@Keab42

Someone I work with prefaced a draft memo with "Raising the bar requires us to think boldly about what's possible. It means we must confront assumptions and commit to evolving together."

My first thought on reading this was: Please, tell me this is written by AI.

My second thought was: I am not adding my name to that.

Happy they agreed to slash the bullshit and get to the point.

=> More informations about this toot | More toots from PapyrusBrigade@mstdn.ca

Written by Matt Tearle on 2024-10-29 at 01:45

@PapyrusBrigade @Keab42

I would sign off on that if the entire rest of it was “This week’s team meeting is moved from 9:00 to 9:15”

=> More informations about this toot | More toots from mtearle@universeodon.com

Written by Strypey on 2024-10-27 at 22:40

@Keab42

we should take a look at how we communicate rather than boiling the oceans to have LLMs hallucinate at each other on our behalf

But a Trained #MOLE can write a boring, boilerplate nag email that nobody reads despite it being sent as HTML to millions of people, much faster than the human grifter trying to sell you on a Trained MOLE can do it. Sign up for my half-price 90 day trial for only $3000, and see for yourself! ; )

https://disintermedia.substack.com/

=> More informations about this toot | More toots from strypey@mastodon.nzoss.nz

Written by JustAFrog on 2024-10-27 at 23:02

@Keab42 I just wonder why we can't do terse messages, if that's what we all want.

It's ludicrous to put in lots of words just because everyone puts in lots of words.

=> More informations about this toot | More toots from justafrog@mstdn.social

Written by it hoots for thee on 2024-10-27 at 23:15

@justafrog @Keab42 short and informal is the norm in sweden. like an SMS.

i can't say i saw a point to the longer and more formal style i experienced in california and germany.

your name and email address is already in the headers, i don't need it again, nor the filler stuff that sandwiches the actual message.

=> More informations about this toot | More toots from owl@u8.is

Written by BlackPixelDust: Art :autism: on 2024-10-27 at 23:26

@Keab42 Ditto with writing reports and essays?

I've read so many that take so many paragraphs to say one thing, or they say nothing at all. It's like some folks are getting paid by the word instead of imparting factual information.

(I'm not talking about writing literature. But literature has had that problem long before the digital age.)

=> More informations about this toot | More toots from jtphillipsmnr@mastodon.art

Written by FinalOverdrive on 2024-10-28 at 01:20

@Keab42 i guess we could just read and write the email.

=> More informations about this toot | More toots from FinalOverdrive@kolektiva.social

Written by Kevin Karhan :verified: on 2024-10-28 at 02:13

@Keab42 +9001%

That's why I declined to permit AI use on my CV by HR firms because if they can't be assed to read my stuff and do a 5min call with questions on their part,

then why should they get paid 8⅓-10% of my gross annual paycheck as commission (to be paid by the firm that would hire me through them)?

=> More informations about this toot | More toots from kkarhan@infosec.space

Written by ScriptFanix ❤️ ⏚ ⸫ on 2024-10-28 at 23:24

@Keab42

Fuck LLMs 🖕

@eowyn

=> More informations about this toot | More toots from ScriptFanix@maly.io

Written by ScriptFanix ❤️ ⏚ ⸫ on 2024-10-28 at 23:27

@Keab42

I once saw a PAYING HAProxy Enterprise customer use freaking chatgpt during a call, when ME, a HAProxy support engineer, was just there! WTF has happened to us?

@eowyn

=> More informations about this toot | More toots from ScriptFanix@maly.io

Written by ScriptFanix ❤️ ⏚ ⸫ on 2024-10-28 at 23:29

@Keab42

Honestly, I felt insulted. Like... It's an AI better than the human you're currently talking to?

@eowyn

=> More informations about this toot | More toots from ScriptFanix@maly.io

Written by Vicky Rampin (née Steeves) on 2024-10-29 at 01:28

@Keab42 i work at a uni and i've been saying soon it'll just be bots grading bots and the work is for naught

=> More informations about this toot | More toots from VickyRampin@octodon.social

Written by David Chisnall (Now with 50% more sarcasm!) on 2024-10-29 at 07:41

@Keab42 A few months ago, someone explained to me how they use GitHub Copilot. They didn’t use it as a tool for autocompletion, they used it to identify missing abstractions. If Copilot could autocomplete a block of code, it was a sign that they needed to provide APIs that required less boilerplate.

I think this is a fairly good approach to generalise: LLMs can be used to identify poor processes. If an LLM can do a task well, that task probably doesn’t need doing.

In a few cases, there are cultural reasons why an LLM as a patch is a better fix. For example, LLMs are good at translation and using them is probably easier than making everyone speak Esperanto.

=> More informations about this toot | More toots from david_chisnall@infosec.exchange

Written by kcarruthers on 2024-11-01 at 08:38

@Keab42 yes!!!!

=> More informations about this toot | More toots from kcarruthers@mastodon.social