Anybody who has worked in IT support in the trenches in enterprise IT will tell you there are some Excel power users who basically run the company, are macros wizards and actual ninjas.. about 0.1% of the workforce.
-
@bontchev @GossiTheDog Why would your hypothetical employee have become annoyed with you?
@xan @bontchev @GossiTheDog says more about the poster than the benefits of AI, IMO.
-
I’m really serious about this one btw. Companies have no measurable way of knowing what employees are doing with GenAI. They’re giving Claude Code out like it’s candy and just presuming everybody is an IT power user. They aren’t. They’re converting PDFs and vibe coding garden planning tools.
Copilot M365 has a fake dashboard showing how productive people are.. it has no actual data. It just shows people use it. It’s CIO porn for the CEO. Orgs are pissing money up a wall worldwide.
@GossiTheDog And then there is the non-monetary costs. We recently completed our first penetration test against Copilot in my corp and to say we found a lot of secret and confidential stuff out there just for the prompting is an understatement. The company totally believed Microsoft when they said everything would be safe guarded, yet I personally found a document with every marketing service account name and password. Vender contracts, company secrets, legal documents: we found it all. Copilot in a corporate environment is the single largest amplifier of poor IAM configurations. It is the largest insider threat I’ve ever seen.
-
I’m really serious about this one btw. Companies have no measurable way of knowing what employees are doing with GenAI. They’re giving Claude Code out like it’s candy and just presuming everybody is an IT power user. They aren’t. They’re converting PDFs and vibe coding garden planning tools.
Copilot M365 has a fake dashboard showing how productive people are.. it has no actual data. It just shows people use it. It’s CIO porn for the CEO. Orgs are pissing money up a wall worldwide.
Yes and people post on mastodon in their work supplied browsers too.
If your manager doesn't understand whether you do good work or not, it's natural to exploit that. That's not a tool problem, it's a manager problem!
-
@GossiTheDog honestly, isn't most of the stuff IT teams do nowadays extremely wasteful?
I look at CI pipelines and feel the need to scream.
Upgrade your OS image with hundreds of packages on every push,
Build all layers of your container every time...
Then migrate to the next git service and CI/CD framework every year, complete rewrites.
How many bloody Artifactory mirrors does a company need?!!!Etc. etc. These kids should start with a C64 or ZX80 before let loose on this hot garbage
@brnrd @GossiTheDog There is a ‘certain amount of waste’ built into the modern way of writing, testing and deploying software.
That amount can be totally insane when changing a minor detail. If the pipeline is not able to isolate what to build, test and deploy it will just do it all. Is it necessary? No. Is it able to prevent mistakes? Maybe.
-
@GossiTheDog And then there is the non-monetary costs. We recently completed our first penetration test against Copilot in my corp and to say we found a lot of secret and confidential stuff out there just for the prompting is an understatement. The company totally believed Microsoft when they said everything would be safe guarded, yet I personally found a document with every marketing service account name and password. Vender contracts, company secrets, legal documents: we found it all. Copilot in a corporate environment is the single largest amplifier of poor IAM configurations. It is the largest insider threat I’ve ever seen.
@GossiTheDog PS: we did raise these concerns when management announced everyone was getting Copilot. Took months (and months) to get them to agree to a test. “No one will be able to see anything they don’t already have access to,” they said.
️Yes. Exactly.
-
In an era where companies need to become more efficient and diverse they’ve basically picked the least efficient way to do it, with the biggest risks and highest costs - because everybody else is doing it.
I know somebody at one of the big 4 who has written something in Claude that prompts Claude each twenty minutes for a question, then feeds Claude’s question back into Claude to use their tokens - because token usage is factored into employee evaluations. What are we even doing.
@GossiTheDog I’m starting to see some companies log all AI prompts and review them to see what people are using AI to do. The ones I’ve seen do this are positioning it as a way to discover additional training opportunities for their staff — to train them that they gave other, better tools to compare PDFs, for example. It makes sense if you have the tools and resources to put together this kind of analysis.
-
RE: https://mastodon.social/@404mediaco/116908074107231828
Anybody who has worked in IT support in the trenches in enterprise IT will tell you there are some Excel power users who basically run the company, are macros wizards and actual ninjas.. about 0.1% of the workforce. About 99% of people can’t align a table in Word.
Giving the 99% of people tools which cost $$$ per user a month and letting them do anything is like giving a child a car, and being surprised when they ram the car into a wall three days later and cost $10k after achieving nothing.
@GossiTheDog Fucking. Bingo.
-
In an era where companies need to become more efficient and diverse they’ve basically picked the least efficient way to do it, with the biggest risks and highest costs - because everybody else is doing it.
I know somebody at one of the big 4 who has written something in Claude that prompts Claude each twenty minutes for a question, then feeds Claude’s question back into Claude to use their tokens - because token usage is factored into employee evaluations. What are we even doing.
@GossiTheDog I would add to yours, that it's a good way to do things in the least diverse way possible, as they are all pivoting everything to a common point: the statistical standard AI is trained on. Everything will look the same.
-
@brnrd @GossiTheDog There is a ‘certain amount of waste’ built into the modern way of writing, testing and deploying software.
That amount can be totally insane when changing a minor detail. If the pipeline is not able to isolate what to build, test and deploy it will just do it all. Is it necessary? No. Is it able to prevent mistakes? Maybe.
@jeroenvanbergen @GossiTheDog
Some waste is inherent. But frugality is nowhere to be found nowadays.The FreeBSD ports I maintain I will build / test on all tier-1 platforms I support.
You need to build and run the test-suite on whatever changes you provide.Shit gets out of hand quickly.
We needed to migrate BitBucket to a cluster because of the load on the system, but couldn't be arsed to punish people doing full git clones continuously instead of restricting depth and cloning only the branch you need.
I see pipelines doing the same builds multiple times for different purposes, why?Convenience not only trumps security, it also trumps efficiency.
(yes, I know how awful bitbucket is, don't @ me)
-
@GossiTheDog honestly, isn't most of the stuff IT teams do nowadays extremely wasteful?
I look at CI pipelines and feel the need to scream.
Upgrade your OS image with hundreds of packages on every push,
Build all layers of your container every time...
Then migrate to the next git service and CI/CD framework every year, complete rewrites.
How many bloody Artifactory mirrors does a company need?!!!Etc. etc. These kids should start with a C64 or ZX80 before let loose on this hot garbage
@brnrd @GossiTheDog it doesn't have to be if you're considering those issues you've pointed out worth addressing. Though you're probably right that plenty don't. You could build small purpose-built container images that have the required tools installed and cache them on the runners. You can have proxy registeries / pull through caches to not download the same npm/pypi/maven/... packages all over again.
-
@jeroenvanbergen @GossiTheDog
Some waste is inherent. But frugality is nowhere to be found nowadays.The FreeBSD ports I maintain I will build / test on all tier-1 platforms I support.
You need to build and run the test-suite on whatever changes you provide.Shit gets out of hand quickly.
We needed to migrate BitBucket to a cluster because of the load on the system, but couldn't be arsed to punish people doing full git clones continuously instead of restricting depth and cloning only the branch you need.
I see pipelines doing the same builds multiple times for different purposes, why?Convenience not only trumps security, it also trumps efficiency.
(yes, I know how awful bitbucket is, don't @ me)
@brnrd @GossiTheDog There is a certain irony to burning so many CPU cycles to make minor steps forward in quality or features.
For software that is run by a lot of machines that might be fine, but most corporate software uses the same kind of pipeline these days. The ratio of effort to build, test and deploy that kind of software vs the times a feature is actually used could be wrong for a long time.
-
@brnrd @GossiTheDog it doesn't have to be if you're considering those issues you've pointed out worth addressing. Though you're probably right that plenty don't. You could build small purpose-built container images that have the required tools installed and cache them on the runners. You can have proxy registeries / pull through caches to not download the same npm/pypi/maven/... packages all over again.
@brnrd @GossiTheDog Don't stuff everything in platform specific CI tools, but rather have them in scripts. You can thus test them properly and also have an easier time if CI systems should change for whatever reason. Similarly, have local mirror of git repositories if you're building from source. All of that is unglamorous work though and doesn't generate revenue, so I suspect that many companies don't give engineers much time and resources to address this unless they see releases at risk because jobs get too slow or fail.
-
I’m really serious about this one btw. Companies have no measurable way of knowing what employees are doing with GenAI. They’re giving Claude Code out like it’s candy and just presuming everybody is an IT power user. They aren’t. They’re converting PDFs and vibe coding garden planning tools.
Copilot M365 has a fake dashboard showing how productive people are.. it has no actual data. It just shows people use it. It’s CIO porn for the CEO. Orgs are pissing money up a wall worldwide.
@GossiTheDog good let the. When the bubble bursts they’ll be held accountable by shareholders.
-
In an era where companies need to become more efficient and diverse they’ve basically picked the least efficient way to do it, with the biggest risks and highest costs - because everybody else is doing it.
I know somebody at one of the big 4 who has written something in Claude that prompts Claude each twenty minutes for a question, then feeds Claude’s question back into Claude to use their tokens - because token usage is factored into employee evaluations. What are we even doing.
@GossiTheDog this autofeeding also happens at facebook for the same reason. Tokenmaxing.
-
@GossiTheDog This is why I very seldomly use cloud AI and on purpose try to do everything with a local LLM setup. We know the subsidized cloud plans are going to come crumbling down so it's better to not have becomed dependent on them in your workflow.
@troed @GossiTheDog
A friend of mine doing a lot of leisure picture generation also uses local engines.
He is a gamer. And when not gaming the GPU can be used for the pictures stuff. -
@GossiTheDog I wish I could find the original to attribute it properly but this fits
@malwareminigun @GossiTheDog No attribution necessary, it's AI generated.
-
@GossiTheDog I wish I could find the original to attribute it properly but this fits
@malwareminigun @GossiTheDog ironically i think this image might be ai?
-
RE: https://mastodon.social/@404mediaco/116908074107231828
Anybody who has worked in IT support in the trenches in enterprise IT will tell you there are some Excel power users who basically run the company, are macros wizards and actual ninjas.. about 0.1% of the workforce. About 99% of people can’t align a table in Word.
Giving the 99% of people tools which cost $$$ per user a month and letting them do anything is like giving a child a car, and being surprised when they ram the car into a wall three days later and cost $10k after achieving nothing.
@GossiTheDog ar my first job, we fucked up CSRF integration, it blocked paypal from confirming payment from client. The #2 in the enterprise noticed after a couple hours. She asked for the raw list of all order id, she plucked all paypal payments, put both list in excel, applied some magic and got 4 impacted order. By the time I opened a shell to the DB, she had basically fixed the issue.
-
I’m really serious about this one btw. Companies have no measurable way of knowing what employees are doing with GenAI. They’re giving Claude Code out like it’s candy and just presuming everybody is an IT power user. They aren’t. They’re converting PDFs and vibe coding garden planning tools.
Copilot M365 has a fake dashboard showing how productive people are.. it has no actual data. It just shows people use it. It’s CIO porn for the CEO. Orgs are pissing money up a wall worldwide.
You should post this rant over on LinkedIn where the audience that apparently doesn't know this already is.
When you post it here on Fedi you're preaching to the choir.
-
RE: https://mastodon.social/@404mediaco/116908074107231828
Anybody who has worked in IT support in the trenches in enterprise IT will tell you there are some Excel power users who basically run the company, are macros wizards and actual ninjas.. about 0.1% of the workforce. About 99% of people can’t align a table in Word.
Giving the 99% of people tools which cost $$$ per user a month and letting them do anything is like giving a child a car, and being surprised when they ram the car into a wall three days later and cost $10k after achieving nothing.
@GossiTheDog
One thing that has occurred to me is how much people use it for search, where search is (still, if used correctly) broadly factual and free at the point of use, and the big GenAI providers tend to provide affirmatory answers while burning through compute.
