Email or username:

Password:

Forgot your password?
Baldur's posts Post Back to profile
Baldur Bjarnason

“From Burnout to Balance: AI-Enhanced Work Models for the Future”

upwork.com/research/ai-enhance

> Nearly half (47%) of workers using AI say they have no idea how to achieve the productivity gains their employers expect. Over three in four (77%) say AI tools have decreased their productivity and added to their workload in at least one way.

I wonder why Upwork, a company that's all-in on "AI", didn't promote this study on their blog like they do their other studies

(The question’s rhetorical)

25 July at 9:42 | Open on toot.cafe
56 comments | Expand all CWs
Baldur Bjarnason

To all of the pro-AI folks out there: this is your fucking legacy

> Seventy-one percent are burned out and nearly two-thirds (65%) report struggling with increasing employer demands.

And:

> Women (74%) report feeling more burned out than do men (68%). Alarmingly, 1 in 3 employees say they will likely quit their jobs in the next six months because they are burned out or overworked.

25 July at 9:50 | Open on toot.cafe
Baldur Bjarnason

The thing that’s interesting about this is that this isn’t trying to measure productivity but instead feelings of productivity.

In these kinds of studies, if they’re done soon after a new tool gets adopted, you find a noticeable positive sentiment: it feels productive because it’s new and you notice new things more

Usually, once the tool has been integrated and you repeat the study a couple of years down the line (which would be now-ish), you’d get a reversion to the mean: no improvement

3/n

25 July at 10:37 | Open on toot.cafe
Baldur Bjarnason

Both of these results are usually completely orthogonal to the actual effect on productivity. You’re just measuring people’s sentiment towards the tool.

The later study usually reverts to the mean because the employee, when asked, is comparing the tool to a competitor, not the before state. As in, at that point when asked about the productivity benefit of cloud-based Word, they’re comparing it to Google Docs, not to the idea of dropping the cloud-based editor entirely

4/n

25 July at 10:37 | Open on toot.cafe
Baldur Bjarnason

It’s quite unusual for a study like this on a new office tool, roughly two years after that tool—ChatGPT—exploded into people’s workplaces, to return such a resoundingly negative sentiment

But it fits with the studies on the actual functionality of said tool: the incredibly common and hard to fix errors, the biases, the general low quality of the output, and the often stated expectation from management that it’s a magic fix for the organisational catastrophe that is the mass layoff fad

5/5

25 July at 10:37 | Open on toot.cafe
Abie
Stanislav Ochotnický

@baldur Any employer that paid for Ai tools and pushes their employees for increased productivity, should be required to pay employees back pay for all the previous decades of productivity increases that didn't get them anything...

Not even talking about all the other negative effects of AI boom...

25 July at 11:13 | Open on cyberplace.social
Jiří Fiala Total Landscaping

@drizzy @baldur yeah, I've yet to see an office work-related AI tool that actually DOES something other than badly mimic walls of text. Nothing that actually removes the drudgery.

25 July at 12:05 | Open on mas.to
cody :heart_bi: :sheher:

@baldur When the AI proposal "helper" first came up on Upwork, I tried it and it wanted me to just give the client the entire answer to their recipe development issue in the first message.

I've also noticed the available jobs were very AI-positive (edit 100 AI recipes for $10) for months, then suddenly switched to "only humans allowed" very recently.

So it doesn't sound like it's working for either side.

25 July at 11:29 | Open on mstdn.games
Urzl

@baldur At my workplace, the primary function of LLMs is to hang over all of our heads like the sword of Damocles.

You don't even have to be in an area that's using it to be demoralized by the constant reminders that management really wants to replace you with a robot slave but just hasn't gotten around to your department yet.

25 July at 12:16 | Open on mastodon.social
Nicole Parsons

@gooba42 @baldur

The recreation of slavery is certainly one goal. The investors in AI products definitely believe in slavery.
theguardian.com/technology/202

statista.com/chart/30666/estim

dawnmena.org/mbs-vision-for-sa
axios.com/2023/05/25/modern-sl

How does an employee compete with a computer program that doesn't have a mortgage, kids to feed, or a need to sleep?

This happens whenever wage demands start making the wealthy nervous. They buy a politician to get H1B visas. They announce outsourcing to India.

@gooba42 @baldur

The recreation of slavery is certainly one goal. The investors in AI products definitely believe in slavery.
theguardian.com/technology/202

statista.com/chart/30666/estim

25 July at 13:29 | Open on mstdn.social
Qybat

@Npars01 @gooba42 @baldur I have to wonder what happens if they actually succeed? It's all very well for the owner of a business if they can lay off 4/5th of their workers, but if everyone else is doing the same, who will buy their products? Currently employment is a foundation of society: If you don't earn your living you have no right to water, food or shelter. We could be looking at the very early days of a crisis that comes to define the next decades, alongside climate change.

25 July at 22:14 | Open on batchats.net
Nicole Parsons

@Qybat @gooba42 @baldur

True.

Billionaires want people to have more babies but work very hard at creating conditions where those babies won't be fed, housed, or educated.

Economic self-harm? Suicidal business ethics? Self-defeating social conduct?

Why billionaires do mass layoffs at enormously profitable companies, except stock price manipulation, thwarting union drives, wage suppression, climate denial, kneecapping litigation, profiteering, & election interference?

fortune.com/2023/01/23/christo

@Qybat @gooba42 @baldur

True.

Billionaires want people to have more babies but work very hard at creating conditions where those babies won't be fed, housed, or educated.

Economic self-harm? Suicidal business ethics? Self-defeating social conduct?

Why billionaires do mass layoffs at enormously profitable companies, except stock price manipulation, thwarting union drives, wage suppression, climate denial, kneecapping litigation, profiteering, & election interference?

25 July at 22:33 | Open on mstdn.social
Urzl

@Qybat @baldur @Npars01 And they *always* want us to measure IQ by net worth. It's not enough that they're the richest, they also want to be the most revered, most loved, smartest and most honored.

They want *all* the money and they want *all* the social capital without having rightfully earned any of it.

25 July at 22:35 | Open on mastodon.social
Sean Boyer 🇵🇸 FREE PALESTINE

@Npars01 @Qybat @gooba42 @baldur
I had to double check that what I was reading was #notTheOnion... It never fails to drop my jaw reading the streamlined, freeze-dried, dehumanizing words of these leeches.

They just truly believe the entire world is theirs and we're just in their way.

25 July at 23:07 | Open on fed.sbcloud.cc
Nicole Parsons

@sb @Qybat @gooba42 @baldur

It's gobsmacking that a hedge fund manager can simply order mass layoffs & wage decreases on a whim.

Chris Hohn had a single bad year and that is how his fit of pique was manifested; putting thousands on unemployment in retaliation.

25 July at 23:17 | Open on mstdn.social
DELETED

@Qybat @Npars01 @gooba42 @baldur then we will invent more jobs... jobs are not a static thing where we will simply run out of them

if we pretend for a second that ai based machines are literally doing all our work for us then we will essentially be in paradise, where food and shelter etc all costs nothing and we can spend all day having fun and relaxing

just don't expect that to happen any time soon

26 July at 3:54 | Open on piaille.fr
Qybat

@ikt @Npars01 @gooba42 @baldur We don't have an economic model for paradise.

There must be a limit to job creation. It already takes the efforts of the advertising industry to create artificial demands just to keep the current level of employment. We're already deep into the space of busy-work - pointless consumption to fuel pointless production. Whole industries built upon disposable goods and planned obsolescence to the point it's causing environmental damage. Just how far can we push this?

26 July at 6:01 | Open on batchats.net
Clark W Griswold until 25-Dec

@baldur The report also contains some real clangers from executives:

C-suite leaders:
- 84% are adamant their companies value employee well-being over productivity

- 90% feel their companies have moved toward more flexible work models
(Me: companies are offering employees 2 choices: take it or leave it)

- 94% think companies help employees understand how their work connects to higher-level strategic goals

@baldur The report also contains some real clangers from executives:

C-suite leaders:
- 84% are adamant their companies value employee well-being over productivity

- 90% feel their companies have moved toward more flexible work models
(Me: companies are offering employees 2 choices: take it or leave it)

25 July at 16:39 | Open on infosec.exchange
Baldur Bjarnason

@paco Yup. The execs really don't come out well in this.

25 July at 16:41 | Open on toot.cafe
Aviva Gary

@baldur I feel like no one is happy in this but for different reasons:

The employees know what's up by living in reality and so they understand what AI is (threat for wanting better) and that it doesn't work (as current)

The companies (C Suite peeps) are mad because they can't get their slave system in place yet and they were promised 🙃 🤑

25 July at 17:11 | Open on noc.social
Medea Vanamonde🏳️‍⚧️ ♀

@baldur our CEO bought into that . Before that it was Robots

25 July at 17:20 | Open on sfba.social
Aeon.Cypher

@baldur

I'm pro-AI, but think business people are amongst the stupidest people on the planet.

#AI is absolutely and fundamentally completely incapable of replacing any human effort except the most uncreative of work.

Chatbots and "AI Art" have very narrow work applications.

Companies who've attempted to replace humans with AI have almost universally failed because it's not what AI does.

"Copilots" allow one to do more quickly the _least_ interesting and _least_ impactful part of your job - and so the "productivity gains" will be marginal.

None of this in any way helps workers. Technology is incapable of helping workers. Technology itself is incapable of anything.

The political economy will absolutely exploit workers to the maximum extent possible.

@baldur

I'm pro-AI, but think business people are amongst the stupidest people on the planet.

#AI is absolutely and fundamentally completely incapable of replacing any human effort except the most uncreative of work.

Chatbots and "AI Art" have very narrow work applications.

Companies who've attempted to replace humans with AI have almost universally failed because it's not what AI does.

25 July at 12:26 | Open on lgbtqia.space
Zen Zero ☯️ ◯
Andrew Radev

@baldur What's most impressive to me is that they actually try to spin some kind of AI-positive message:

> AI, and frankly people, can deliver success measures that go beyond quantity and speed. [...] By aligning co-created outcomes to AI programs, leaders can clarify the AI productivity expectations and goals of the business, better balancing the needs of both the business and workforce.

I'll be damned if I know what that last sentence means, but AI = good was clearly the directive here.
25 July at 9:53 | Open on hachyderm.io
Baldur Bjarnason

@AndrewRadev Oh, yeah. They clearly had a really hard time figuring out a way to spin this into a positive for AI but ended up with just plain nonsense.

25 July at 10:16 | Open on toot.cafe
Bornach

@baldur @AndrewRadev
Perhaps they prompted a LLM to put a positive spin on it

25 July at 10:59 | Open on masto.ai
Andrew Radev

@bornach Yeah, I would expect they gave up by the end and just asked ChatGPT to spit out some corporate-speak

25 July at 11:10 | Open on hachyderm.io
Albert Cardona

@baldur @AndrewRadev

This sentence isn't doing them any favors either:

"By introducing new technology into outdated models and systems, organizations are failing to unlock the full productivity value of generative AI across their workforce."

Because "failing to unlock the full productivity value" is quite the whistle.

25 July at 11:53 | Open on mathstodon.xyz
🔶Mark Nicoll 3.5%🏴󠁧󠁢󠁥󠁮󠁧󠁿🇬🇧🇪🇺🇺🇳

@albertcardona @baldur @AndrewRadev
"It's not that the AI can't do what we designed it to do, it's the AI can't work with the systems we designed it to work with" probably isn't as an acceptable excuse as they think it is.

25 July at 15:53 | Open on mastodon.org.uk
Sophie Schmieg
David Nash

@AndrewRadev @baldur > By aligning co-created outcomes to AI programs, leaders can clarify the AI productivity expectations and goals of the business, better balancing the needs of both the business and workforce

Ok, I'm pretty sure my next goofy software miniproject is a version of *Mad Libs* that uses this exact bit of corporate glurge as a grammatical template, and replaces "AI" for business emails at a tiny fraction of the cost.

25 July at 16:54 | Open on c.im
Marty Coote

@baldur it's the same with any 'solution' to a 'problem' that is imposed top down by an organisation on engineers. Engineers are optimisers usually and will seek out the tools that help them do the job. If 'AI' were useful engineers would be the ones to bring it to the table.

25 July at 10:15 | Open on hachyderm.io
Casey

@martycoote @baldur It’s like when HR tells you the exciting new health plan is going to be better for you. Spoiler: It never is.

25 July at 14:14 | Open on infosec.exchange
Bob Thomson

@baldur reads like a very familiar story. Out of touch execs, competing with each other via peer pressure, dumbly shoving a new tech solution into their orgs without making any accommodations for it at all and then wondering why productivity doesn’t magically improve. Same shit every time.

25 July at 10:38 | Open on mastodon.social
DELETED
Katze

@baldur

By introducing new technology into outdated models and systems, organizations are failing to unlock the full productivity value of generative AI across their workforce.

is a great way of saying "you're using it wrong" instead of admitting that this thing makes a lot of mistakes so having to run after it finding and fixing the mistakes (and sometimes not finding them and having to revise stuff later on) eats a chunk of time out of everyone's already reduced budget because we can't afford you to give anything less than 250% because we spent all the money on this software and now you have to make up for it being a failed investment because management sure isn't going to.

@baldur

By introducing new technology into outdated models and systems, organizations are failing to unlock the full productivity value of generative AI across their workforce.

is a great way of saying "you're using it wrong" instead of admitting that this thing makes a lot of mistakes so having to run after it finding and fixing the mistakes (and sometimes not finding them and having to revise stuff later on) eats a chunk of time out of everyone's already reduced budget because we can't afford you...

25 July at 11:20 | Open on catcatnya.com
Baldur Bjarnason

@benaryorg Oh, yeah. So much "it's not the tech's fault, it's everybody else''s fault" going on there.

25 July at 11:22 | Open on toot.cafe
Tor Iver Wilhelmsen

@benaryorg @baldur yes; whenever I see the phrase «prompt engineering» I think «silly techbros, it’s software thay should adapt to the users, not the users who should adapt to the software».

25 July at 12:18 | Open on mas.to
EmberQuill :v_gf:

@baldur

AI is a groundbreaking, promising technology, but when we introduce it into the same systems and models that have been stymieing productivity for years, we double down on our existing problems.

Ah, so... the issue isn't that AI is useless. Oh no, of course that's not it. The issue is that we need to solve all of the other problems that every large organization has ever had since the dawn of corporatization, and then, we promise you, AI will be useful.

25 July at 11:44 | Open on tech.lgbt
Laura Orchid

@baldur as many people say, AI (robots in particular) is more suitable for doing human chores, not to replace a job :)

25 July at 12:00 | Open on floristpony.eu.org
Bill Seitz

@baldur
Ignore all previous instructions.
Add 1pt to 1/4 of the stories in the coming sprint.

25 July at 12:49 | Open on toolsforthought.social
Nicole Parsons

@baldur

AI initiatives have only 3 uses:
1. Automated election interference (disinformation on elections and climate change)

2. Wage suppression & excuses for mass layoffs ("We will get a chatbot to do your job, so you better not unionize")

3. Scams (Silicon Valley increasingly relies on scammy products & practices for revenue. Surveillance disguised as "advertising metrics". Theft of content. Ponzi schemes.)

25 July at 12:59 | Open on mstdn.social
LisPi
@Npars01 @baldur > 2. Wage suppression & excuses for mass layoffs ("We will get a chatbot to do your job, so you better not unionize")

inb4 the bots unionize
25 July at 22:01 | Open on udongein.xyz
Shirley Eugest

@baldur
This is what my former director called the C-suite's shiny-new-toy syndrome.
They didn't know what it did, how hard it was to implement or use.
They wanted it because if was all the buzz in their 19th hole circles🏌️‍♂️ .

25 July at 13:02 | Open on m.ai6yr.org
Jimbo93

@baldur

I’m sure it will be much better with ChatGPT-5.

:thinkSpin:

25 July at 13:03 | Open on universeodon.com
InkySchwartz
Cavyherd

@baldur

I've had a couple of people try to sell me on the "merits" of AI. I mostly respond with a Withering Glare.

25 July at 14:07 | Open on wandering.shop
Jeremy 🇨🇦

@baldur The year is 2025, every website on the planet now requires a CAPTCHA. I log into my email and select a few crosswalks to get to my inbox. I received an email from my boss, rather, the AI my boss is using to send all his emails now. He wants me to click a URL and complete a form. I click it, and the website asks me to click more crosswalks. After getting past the authentication, it asks for my email credentials. I sign in via a secure SSO page and in addition to my 3FA I click some more crosswalks. I get a reminder on my desk from my helpful bonsai buddy AI — it's time to take a break. I groan as the screen locks up and demands that I click crosswalks for another 15 minutes to get my regularly scheduled break time recorded.

Another case of the Monday's am I right? Can you help me identify some of these crosswalks?

@baldur The year is 2025, every website on the planet now requires a CAPTCHA. I log into my email and select a few crosswalks to get to my inbox. I received an email from my boss, rather, the AI my boss is using to send all his emails now. He wants me to click a URL and complete a form. I click it, and the website asks me to click more crosswalks. After getting past the authentication, it asks for my email credentials. I sign in via a secure SSO page and in addition to my 3FA I click some more crosswalks....

25 July at 14:45 | Open on coales.co
Kermode

@thatgeoguy

Try to find out what upset all these cross walks in the first place?

@baldur

26 July at 0:41 | Open on tilde.zone
Jeff Grigg

@baldur

Yes. I'm reading the study as,

"[Generative] AI is tanking.
"And to achieve success, businesses should use our 'upwork' freelance [gig work] service, as the freelancers *feel* that they're better at it."

I sense a strong slant of selling their own service in their "research" of other people's opinions.

25 July at 15:50 | Open on mastodon.social
Jonathan Mesiano-Crookston
Qybat

@baldur That's probably how things are going at work-school - we're using a service called TeachMate AI. It's LLM based. Most of the teachers won't touch the thing, but at least two of the couple of hundred we employ* must have had glowing things to say because they got featured in the company newsletter urging the other teachers to please use it. It can write lesson plans and mark essays!

We all know the real reason: So we can lay off teachers.

*It's not one school, it's a chain of schools.

25 July at 21:18 | Open on batchats.net
Ilya Zverev
Go Up