ChatGPT (AI)

[Digsey]

Nothings wrong with that, but why would you not streamline something if you can?

The way it works is I get it to generate a 4 day meal plan with my prompt of:

- low in cholesterol
- easy to make within 30 mins or so
- can be bought at Aldi or Morrisons

And then every week it will generate a different meal incorporating different meats into it for variety, essentially a hello fresh or gousto meal plan without paying through the roof for it.

Then it will:

- generate the meal plan
- give me the shopping list in order of the shop floor plan.
- then create a PDF, one page per meal with instructions and Qty hello fresh style that I will simply open up when cooking

I can't see how that cant possibly make someone's life easier*

*when it doesn't chuck f**king feta in when not needed.

is there a way to integrate this into an online shopping list, for say tesco? so it opens a basket and adds the required items?

this seems like a cool thing to do without, as you say, paying for gusto etc, but adding them all to a basket would make this perfect!
Can it also remember previous orders so you dont double buy everything every week?

 

[dann2707]

is there a way to integrate this into an online shopping list, for say tesco? so it opens a basket and adds the required items?

this seems like a cool thing to do without, as you say, paying for gusto etc, but adding them all to a basket would make this perfect!
Can it also remember previous orders so you dont double buy everything every week?

That would be amazing but I am not too sure being honest mate!

I have it set up so that I just tell it to put on the list the right Qty for that week so we have zero waste like gousto.

And sometimes if we have some bits left over from other meals at the weekend or whatever I'll tell it I have that in stock and it will incorporate into my meal plan.

And also sometimes it can get a bit repetitive so I get it to list out 15 ideas and I pick from them. Works really well when it does

 

[dann2707]

This is what that project looks like within Chatgpt now. I've created it as a side project as "healthy eating plan"

Can see half way down it started to rattle me

 

[Digsey]

haha, all my AI conversations end up with me shouting at it.

I'll have a look at doing something similar, we've always got loads of food in, but sometimes i run out of ideas of what to do with it, might try feeding it my tesco orders and seeing if i can create meals with what i've bought, the opposite to yours i guess.

Thanks for idea though, i've used Ai for a number of things in the past, but its sometimes more difficult thinking of a task for it, and how it could help.

 

[dann2707]

haha, all my AI conversations end up with me shouting at it.

I'll have a look at doing something similar, we've always got loads of food in, but sometimes i run out of ideas of what to do with it, might try feeding it my tesco orders and seeing if i can create meals with what i've bought, the opposite to yours i guess.

Thanks for idea though, i've used Ai for a number of things in the past, but its sometimes more difficult thinking of a task for it, and how it could help.

That's a good idea!

Yeah f**k you @Flat Eric!!

 

[Rskennedy]

is there a way to integrate this into an online shopping list, for say tesco? so it opens a basket and adds the required items?

this seems like a cool thing to do without, as you say, paying for gusto etc, but adding them all to a basket would make this perfect!
Can it also remember previous orders so you dont double buy everything every week?

Should be able to do that with Claude cowork. I imagine it would be like a toddler though and lob a load of random s**t in your basket as well.

Or do something daft like add 300 lemons instead of 300g of lemons

 

[Digsey]

so exactly the same as when i do it myself then? i've never used claude so i'll have a look, thanks

 

[Adamm.]

All these posts about what AI can do I just hope you guys are using please and thank you

 

[charltjr]

All these posts about what AI can do I just hope you guys are using please and thank you

In all seriousness, if you are then don’t because it increases the cost of the query. Politeness is expensive, who knew

Had some fun with a “but Claude says” person on Friday. I don’t care what Claude says, Claude is wrong because it doesn’t understand X, Y and Z. “But….. Claude says”

 

[Sidnancy]

Off the back of my 'I dont know how to use AI' post, ive started using it for something a bit more serious than 'What should I do in the gym tonight'

Card Machines.

About 3 years ago, I renegotiated our card machines at work, to some unbelievable effect. Saved the company a literal fortune, overnight. Bonza.

So, ive started the process again to renegotiate new terms. The volume of utter shite, combined with technical terms, is amazing. One company offers an amazing rate, but has crazy back end fees. Another looks rubbish based on higher rates, but has zero back end charges so becomes a better option. Ive got 4 or 5 of them all fighting over me right now, every time they offer me something else, I just chuck it into chatgpt.

It doesnt stop the natural negotiation of it all, but its brilliant to get it to compare things against each other. I had another offer come in this morning that looked amazing, but it still wasnt as cheap as my Jeweler mate round the corner who uses them .

Its proper smoke and mirrors, the card machine game!

 

[Matt Cup]

That’s it but not the same guy unfortunately, it’s actually these seats I’m after if it’s easier to find them.

View attachment 1800375

View attachment 1800376

Saw these for sale which reminded me of your post.

 

[Louis]

Saw these for sale which reminded me of your post.

View attachment 1801773

For your lounge @James

 

[MLB]

This is what that project looks like within Chatgpt now. I've created it as a side project as "healthy eating plan"

Can see half way down it started to rattle me

As I started the thread, I feel I should know more. But I genuinely don't, and im not afraid to admit that.

How do you guys work out how to use AI?

Sounds like a silly question, but with things changing so rapidly and new 'features' being developed all the time - how do you keep up?

Im beginning to feel a little bit like my old man with technology, its moving faster than I can keep up with.

I have the paid version of GPT, but purely because it allows me to have the privacy setting on to comply with GDPR. Ill be honest though, I mostly use it for my gym program and calorie counting. That and, ill be honest, I use it to 'fight fire with fire' when customers fire out some insane AI emails at me.

I notice on the PC app, there's a 'Work' tab now. It appears its for integrating into apps such as outlook etc. Whilst that would be handy, in some cases, I think its not for me just yet. I also dont really trust it enough to do some things.

So, help?

I've just asked chatgpt how I should use it. Gave me a pretty decent explanation of how I could make better use of it in projects, keeping track of changes etc.

For example it suggested to have one project for home automation. Use a 'work chat' as the main chat as it will be an ongoing project which needs better/deeper handling and use normal chats for quick on the spot questions.

Then in the work chat I've asked to keep a running log of
Current system
Rollout plan
Change log
Decisions register
Device inventory
Ideas/backlog

Everytime I add a device, or create an automation etc chatgpt automatically adds devices to the system and device inventory, updates the change log of what's done and when and adds the 'why' for certain changes in the change log.

If I get stuck or have an idea that gets dumped in the backlog.

This way I can discuss and build the home automation system while keeping things semi organised.

I do the same with my 'Workout' project. Have a workout log where I just prompt 'ready for a workout' and chatgpt automatically guides me through the session, while automatically updating weights, reps etc. If certain exercises don't fit anymore the 'master plan' where I discuss progress and changes to the routine gets uodated without being clogged up by random workout sessions.

Also makes it easier to make monthly/quarterly progress reviews.

If there are specific requirements of how you would like chatgpt to respond or certain parameters you want to set for that specific project you can add those in the project main view, settings and then 'Edit instructions'.

 

[TC.]

Saw these for sale which reminded me of your post.

View attachment 1801773

Jeez they’re grim, fella I know was after 16 to kit his sprinter out to do golf tours around Ireland. Haven’t heard from him again so he might have got sorted.

 

[J-J]

Off the back of my 'I dont know how to use AI' post, ive started using it for something a bit more serious than 'What should I do in the gym tonight'

Card Machines.

About 3 years ago, I renegotiated our card machines at work, to some unbelievable effect. Saved the company a literal fortune, overnight. Bonza.

So, ive started the process again to renegotiate new terms. The volume of utter shite, combined with technical terms, is amazing. One company offers an amazing rate, but has crazy back end fees. Another looks rubbish based on higher rates, but has zero back end charges so becomes a better option. Ive got 4 or 5 of them all fighting over me right now, every time they offer me something else, I just chuck it into chatgpt.

It doesnt stop the natural negotiation of it all, but its brilliant to get it to compare things against each other. I had another offer come in this morning that looked amazing, but it still wasnt as cheap as my Jeweler mate round the corner who uses them .

Its proper smoke and mirrors, the card machine game!

Yeah it's great for deciphering things where there are different business models in play.

In your game would it not work for generating mailing lists and follow-ups?

Or a pricing programme that automatically generates quotes etc?

 

[ChrisR]

I've been playing about with it lately at work with an agent built around our product documentation to put together demo and presentation tracks which has been quite useful, quick to build something tailored to what I need rather than using boiler plate documents.

But it's been key doing it with the documentation focused agent, else it'll be hallucination city, so it needs a curated set of data to work with.

Also work have an agent for completing your yearly performance reviews which as a new starter has been a god send, it guides you through everything and helps you align everything to the company language/frameworks etc.

 

[Darren S]

The human race with AI = "We're all turkeys, voting in Christmas."

We're in the 'oh wow' phase of what AI can do and riding that wave of enthusiasm without much in the way of second-guessing it. I've used it for work previously, where it's been great for assisting in user story creation and suggesting auxiliary features that I might not have considered myself. The trouble is with most businesses flying full ahead with AI adoption - is that ultimately you don't need as many staff. When the next 100, 1000 and 10,000 businesses all start behaving the same and laying off staff, you're creating a serious economic crisis. There's no point in your own business creating products and services for a population that simply cannot afford them.

I do find it slightly unnerving also, where even in casual things like PC games, a simple screenshot uploaded to ChatGPT can get it to generate a defensive plan in a strategy game. Took 15-20secs for it to respond.

It's fuc*ing Wargames from 1983, yet infinitely more capable....

 

[cs_rich]

The human race with AI = "We're all turkeys, voting in Christmas."

We're in the 'oh wow' phase of what AI can do and riding that wave of enthusiasm without much in the way of second-guessing it. I've used it for work previously, where it's been great for assisting in user story creation and suggesting auxiliary features that I might not have considered myself. The trouble is with most businesses flying full ahead with AI adoption - is that ultimately you don't need as many staff. When the next 100, 1000 and 10,000 businesses all start behaving the same and laying off staff, you're creating a serious economic crisis. There's no point in your own business creating products and services for a population that simply cannot afford them.

I do find it slightly unnerving also, where even in casual things like PC games, a simple screenshot uploaded to ChatGPT can get it to generate a defensive plan in a strategy game. Took 15-20secs for it to respond.

It's fuc*ing Wargames from 1983, yet infinitely more capable....

View attachment 1801844

Humans are notoriously bad at dealing with long-term risks that require behavioural change, e.g. climate change, and now AI. So yeah, we're probably all fucked.

I started my new role in Feb '25 when everything was taking off and I swear that most of my role now feels like acting as an intermediary between my Claude instance, and other staff / customers / suppliers etc. And even more recently it's clear that I'm acting as an intermediary between my Claude, and their AI. If my post wasn't actually required for legal reasons I'd worry I was completely replaceable within the next 12mths.

 

[charltjr]

Humans are notoriously bad at dealing with long-term risks that require behavioural change, e.g. climate change, and now AI. So yeah, we're probably all fucked.

I started my new role in Feb '25 when everything was taking off and I swear that most of my role now feels like acting as an intermediary between my Claude instance, and other staff / customers / suppliers etc. And even more recently it's clear that I'm acting as an intermediary between my Claude, and their AI. If my post wasn't actually required for legal reasons I'd worry I was completely replaceable within the next 12mths.

We’re seeing massive pressure to use AI and one of our support units has been going big on using it for root cause analysis etc. Even with something state of the art for tech work like Claude though, it spits out b****cks when it doesn’t know and what’s so much worse is it sounds utterly confident when doing it.

For infrastructure stuff it’s good. Seriously impressive for solution use - write me a query that does x,y,z or a script to do x,y,z. Absolute and utter bobbins for doing diagnosis on things that haven’t already been stuck into it’s database through whatever the model has been trained on.

The consequences will be coming home to roost soon enough in terms of where it’s used and where it isn’t. But would I want to be reliant on a job looking after some industry standard and very well known infrastructure or products at this point? Not so much.

I think it genuinely is the next industrial revolution in the making, and anyone who thinks that’s a good thing for workers in the short term really needs to go and read some history books.

 

[cs_rich]

We’re seeing massive pressure to use AI and one of our support units has been going big on using it for root cause analysis etc. Even with something state of the art for tech work like Claude though, it spits out b****cks when it doesn’t know and what’s so much worse is it sounds utterly confident when doing it.

For infrastructure stuff it’s good. Seriously impressive for solution use - write me a query that does x,y,z or a script to do x,y,z. Absolute and utter bobbins for doing diagnosis on things that haven’t already been stuck into it’s database through whatever the model has been trained on.

The consequences will be coming home to roost soon enough in terms of where it’s used and where it isn’t. But would I want to be reliant on a job looking after some industry standard and very well known infrastructure or products at this point? Not so much.

I think it genuinely is the next industrial revolution in the making, and anyone who thinks that’s a good thing for workers in the short term really needs to go and read some history books.

Cory Doctorow has a good book on it "The Reverse-Centaurs Guide to Life After AI'. He draws the distinction between people who use AI as a tool and people who have AI applied to them. So at the minute it's a great tool for more senior coders, managers, leadership etc., but it's f**king awful for gig-workers, junior coders, junior solicitors etc.

Problem is that bar is moving all the time and there's no guarantee there'll be jobs for those displaced workers to move into.

 

[charltjr]

Cory Doctorow has a good book on it "The Reverse-Centaurs Guide to Life After AI'. He draws the distinction between people who use AI as a tool and people who have AI applied to them. So at the minute it's a great tool for more senior coders, managers, leadership etc., but it's f**king awful for gig-workers, junior coders, junior solicitors etc.

Problem is that bar is moving all the time and there's no guarantee there'll be jobs for those displaced workers to move into.

Yes, and the other big question is where’s the pipeline of expertise to do the stuff that AI can’t going to come from.

AI isn’t remotely intelligent. Yet. Boil it right down (with apologies to anyone who’s deep into the tech because this is wrong on almost every technical level, but it’s a useful analogy) and It’s a pattern matching engine on steroids.

 

[cs_rich]

Yes, and the other big question is where’s the pipeline of expertise to do the stuff that AI can’t going to come from.

AI isn’t remotely intelligent. Yet. Boil it right down (with apologies to anyone who’s deep into the tech because this is wrong on almost every technical level, but it’s a useful analogy) and It’s a pattern matching engine on steroids.

I'm just a pattern matching engine without the steroids. I'm fucked.

It's interesting though that what we have now is already a long way away from plain old LLM. I see a lot of people banging on about how it's just fancy autocomplete, but just go and ask Claude how much is core LLM and how much is now a suite of multi-modal tools, all strung together by LLM for communication and you can see how far it's come already.

This might not get us to ASI but how many people's jobs require much intelligence, let alone super-intelligence. I reckon 99% of businesses would be improved massively simply by moving their processes to the well understood 'best practice'. You don't require creative thinking for that, just a tool capable of research and pulling that together into a plan, something well within the capability of current AI. And if all businesses did that, I'm sure they would immediately shed a lot of staff.

 

[charltjr]

I'm just a pattern matching engine without the steroids. I'm fucked.

It's interesting though that what we have now is already a long way away from plain old LLM. I see a lot of people banging on about how it's just fancy autocomplete, but just go and ask Claude how much is core LLM and how much is now a suite of multi-modal tools, all strung together by LLM for communication and you can see how far it's come already.

This might not get us to ASI but how many people's jobs require much intelligence, let alone super-intelligence. I reckon 99% of businesses would be improved massively simply by moving their processes to the well understood 'best practice'. You don't require creative thinking for that, just a tool capable of research and pulling that together into a plan, something well within the capability of current AI. And if all businesses did that, I'm sure they would immediately shed a lot of staff.

Yep. And, amusingly, it’s the most senior levels of management who should be among the easiest to replace with AI tools. Bet it won’t happen.

 

[cs_rich]

Yep. And, amusingly, it’s the most senior levels of management who should be among the easiest to replace with AI tools. Bet it won’t happen.

"Ladies and gentlemen of the board, allow me to open this meeting by recommending you replace me with a £100 Claude subscription."

Nope.