OV9

I'm slowly adding more AI to my workflow, it's saving time here and there, and also doing some basic tasks like summarising news stories, doing quick analyses of markets and innovations paths, also I only ate what it told me for 10 days. It did eventually tell me that eating dog food is a great idea, but I did persuade it to do that

Not ChatGPT, but another AI.

The whole thing feels a bit weird, but - blimey.

I am not sure what level I am being fooled on - is it really AI or realz pretending it's AI, or AI of realz pretending it's AI or......

https://www.linkedin.com/posts/liamfall ... 47840-31G-

It looks so genuine and yet is utterly content-free. Which is really the tell.

The rate of improvement whilst expected is still alarming. Must have only been 18 months ago when we were laughing because AI was rendering images of people with 7 fingers and 3 arms.
Kind of feels like we're on the pinnacle of humanity in terms of stuff made by actual people. Going forwards a lot of the creative industry will be increasingly made with AI and as a pretty old school type of guy I don't like that.

Seems that big companies are starting to use AI more, Tesco self serve latest anti theft update. Love that it’s already getting called Tesco VAR

Report of the hearing today about the fake cases cited by a barrister: https://www.lawgazette.co.uk/news/barri ... 87.article

The most worrying thing is that she didn’t realise that the Google results were AI-generated. I must admit I find her evidence somewhat implausible (Safari is not a search engine, for example, indicating that she was somewhat naive or ignorant) but this is going to be the default answer, isn’t it. “Oh, I didn’t know it was AI”. Well fucking double check everything - not just lawyers, everyone, because there’s a new age of misinformation coming.

i had another “salesperson using ChatGPT” issue yesterday to fix. Our client sent through some updated contract amendments (that he had sent through without tracking the changes), i was busy so the salesman didn’t thought he’d “help” by asking chatgpt to compare the docs and summarise the changes and whether they were in our favour.

it came up with around 12 changes with varying levels of impact - i obviously told the sales guy to hold off going back to the client.

i then compared the two docs in Word using the compare function. There were 2 changes, both changes i’d requested. the other 10 changes it had “found” were completely fictitious - at best it was where the client had added a space to fix some formatting.

i can only imagine how fucking stupid i’d/we’d look going back to a client to fight 10 changes they hadn’t made because AI told us to.

I don't know why you'd use AI to do that for you when comparing the two versions in Word is the work of a moment? Even if you could rely on its output, it's just another layer of complexity instead of carrying out the basic task.

Seems to me it's generational thing; young people seem to think AI is in some way the answer to everything.

Jobbo wrote: Thu Jun 05, 2025 8:03 am I don't know why you'd use AI to do that for you when comparing the two versions in Word is the work of a moment? Even if you could rely on its output, it's just another layer of complexity instead of carrying out the basic task.

Seems to me it's generational thing; young people seem to think AI is in some way the answer to everything.

Exactly that. The sales guys tend to use AI to generate a lot of the general wording in our presentation docs, so think it’s the answer to everything.

I was chatting to my son about Uni last week and apparently, lots of his mates used it to write their personal statements

It's definitely becoming more prevalent in the workplace.

I've been told by a friend that their new team member (compliance, uni graduate) admitted in a review that he uses ChatGPT or similar to 'help' with virtually every communication he sends. He saw nothing wrong with it.

It was mentioned to him that there will be occasions - meetings, presentations, decision making, understanding legislation, etc, where he won't be able to ask AI the question and will have to work it out himself. As in, actually understanding the subject matter.

I have used AI (not at work) to help me tweak and tidy up documents that I've already written, but many don't do that and have become lazy and want AI to do it all for them. Engaging brain will be too hard, let alone critical thinking!

I do wonder how long it will be before any role that involves data, analytics, code, numbers, etc, will be offloaded to AI programmes. I find it a pretty scary prospect. The white collar roles will surely face their own crisis as blue collar did with machines.

Yes, every client I've got is encouraging their staff to use whatever AI program they've bought-into, with varying degrees of training being given/offered, let alone any seeming thought as to whether its an appropriate tool for them to use based on role.

We're writing an AI usage policy next week, and I'm involved in the quality/sanity checking of it.

I suspect it's gonna start permissive, and I'm going to be dialling that *right* back.

Beany wrote: Fri Jun 06, 2025 3:20 pm We're writing an AI usage policy next week, and I'm involved in the quality/sanity checking of it.

I suspect it's gonna start permissive, and I'm going to be dialling that *right* back.

I wrote one last year. It was fairly strict to counter the results of a survey we did. It was anonymous to get better data, and the responses were eye-opening. I get that security/privacy/thinking isn't everyone's remit, but blimey.

That's not a bad idea, actually....might 'borrow' that

Feel free. I can also recommend running sessions for people about the dangers of it and what the policy says. I did that too, and it instilled a sense of caution. Covered all the usual things about hallucinations, other people's data, uncontrolled and uncontrollable, sneaky inclusion in existing corp/approved apps, and so on. There's enough content that you can use to give it a fun twist by showcasing the sillier things, then point out the serious stuff.

I also had to draft a policy on AI enabled devices and what was it wasn't permitted in the workplace. It got difficult when we tried to manage AI enabled hearing aids.

I am fascinated by the of the coding AI unicorn Builder’s collapse. The idea that they simply had loads of Indian programmers and were pretending it was AI is utterly hilarious. Also supports the idea that going back a few years all a pitch deck needed to confirm was “massive investment opportunity by leveraging AI technology” and investors would start creaming themselves.

https://archive.is/sivD8

Oh, that's because most of the people sucking up large amounts of VC funding for their 'groundbreaking AI systems' are flat out fucking liars, and a significant proportion of people running tech VC funds are immensely gullible idiots.

They absolutely deserve each other.

See also Amazons 'AI' powered shops with no staff and AI doing all the product checking etc.

Turns out a significant proportion of the actual work was being done by, drum roll, indians being paid fuck all.

https://www.theguardian.com/commentisfr ... technology

Basically, if you see anything claiming to be a revolutionary AI powered game changer, it's very safe to assume that it's bollocks from the off, and wait to be proven otherwise.

Yeah, Amazon are touchy about that one, their PR department rather urgently got onto me for talking about that story on the radio. They made it very clear that the remote workers are there to check and annotate the large volumes of AI-generated data to ensure the AI was being trained correctly. I mean that is necessary to train a model correctly, but PR people make my skin crawl though, and there's obviously a lot of truth in the story, because they weren't threatening me.

My gripe with the amount of funding senselessly poured into scammy AI projects is the sheer quantity of cash that it has devoured which could have absolutely done something positive in the world instead

AI - Anonymous Indians?