All things AI

[Ditto]

I've been using ChatGPT via openAI for a couple of years now. ChatGPT is being rolled out at work via Enterprise Copilot. I'm pushing it hard and at times it gets nowhere near hepling with a challenge. It's constantly failing to fulfil it's promises and apologises profusely when I point them out. With the last challenge I challenged it with, I called it out and suggested perhaps it was not the best tool for the task. I then got the best reply yet, where it basically admitted it's failings and could explain it's limitations. It effectively flipped from trying too hard to achieve what it couldn't, to finally admitting it did not have access to the resources it needed to help with my challenge. It was alsmost as if it had 'grown-up'. So, for those of you that are regularly using ChatGPT, how is it going and does it feel to you that ChatGPT has a personality?

[6Foot2]

When I use ChatGPT, I use it as a sounding board, a starting point for ideas.

 

It’s like having an always-available brainstorming partner, and it can point me in directions that I might not have thought of alone, or maybe not arrived at until a longer period of time had elapsed.

 

Also, if I'm stuck, it can help with "unblocking" a route to more progress.

[Andrew_C]

30 minutes ago, Ditto said:

 how is it going and does it feel to you that ChatGPT has a personality?

"Genuine people personality" or "I'm sorry I can't do that Dave" personality?

[Ditto]

3 minutes ago, Andrew_C said:

"Genuine people personality" or "I'm sorry I can't do that Dave" personality?

Your impression of that, whether it;s one of those 2, something else or whether the idea of a machine having 'people personality' just doesn't make sense.

[Andrew_C]

I think you missed the "cultural references"...

[Ditto]

I certainly did but now my memory is playing tricks and I can't remember if we are Hitch-hiking or in an Odyessy - but on reflection maybe it's both.

[Andrew_C]

Indeed both.

[sigma]

Sam Altman has been announcing today 

 

"In December, as we roll out age-gating more fully and as part of our 'treat adult users like adults' principle, we will allow even more, like erotica for verified adults," Altman concluded - quite the titillating tease. 

Like so many other specifics, there was no mention of how OpenAI intends to handle the creation of erotic content using its chatbot, and Altman's responses in the comments of his X post only suggest users won't be shown it if they don't ask for it.”

 

“How OpenAI intends to stop AI-generated erotic imagery from ending up in the hands of minors, prevent the creation of non-consensual imagery, and tackle other AI porn problems is unknown. Given other recent OpenAI news, it seems monetization may be the driving force for the company, which has yet to disclose profitability.”  

 

https://www.theregister.com/2025/10/14/openai_chatgpt_ai_erotica/

 

Will schools still be endorsing this for teachers?

[Roberto]

I suspect the answer they’ll give you is to use ChatGPT enterprise. We’ve (not a school) done this to allow us to control sign-ups from our domains, to avoid unauthorised use and to avoid our sensitive data being used to train their models.

 

 

[sigma]

1 hour ago, Roberto said:

I suspect the answer they’ll give you is to use ChatGPT enterprise. We’ve (not a school) done this to allow us to control sign-ups from our domains, to avoid unauthorised use and to avoid our sensitive data being used to train their models.

 

 

 

I agree, but schools are always reluctant (resistant), to pay for anything they think they can have (should have?), for free.

[Roberto]

Oh yeah. 100%. But OpenAI hear you. OpenAI don't care.(*)

 

(*) neither do the rest of them, not singling out any vendor as worse than the rest here!

Edited October 15, 2025 by Roberto

[paulkerton]

45 minutes ago, sigma said:

 

I agree, but schools are always reluctant (resistant), to pay for anything they think they can have (should have?), for free.

I was once told something should be free "because it's for the children". I just replied "And yet you take a wage yourself"
That did not go down well. 🤣

[6Foot2]

Link: ChatGPT will soon allow erotica for verified adults, OpenAI boss says - BBC News

 

A new avenue of temptation to go off task for students and a new headache for network support staff.

[6Foot2]

Link: Microsoft introduces its AI virtual assistant character, Mico | AP News

 

Quote

Clippy, the animated paper clip that annoyed Microsoft Office users nearly three decades ago, might have just been ahead of its time.

 

Microsoft introduced a new artificial intelligence character called Mico (pronounced MEE’koh) on Thursday, a floating cartoon face shaped like a blob or flame that will embody the software giant’s Copilot virtual assistant and marks the latest attempt by tech companies to imbue their AI chatbots with more of a personality...

 

[sigma]

https://www.bbc.co.uk/mediacentre/2025/new-ebu-research-ai-assistants-news-content

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory.

[6Foot2]

Link: Armed police handcuff teen after AI mistakes crisp packet for gun in US - BBC News

 

Quote

A US teenager was handcuffed by armed police after an artificial intelligence (AI) system mistakenly said he was carrying a gun - when really he was holding a packet of crisps.

 

"Police showed up, like eight cop cars, and then they all came out with guns pointed at me talking about getting on the ground," 16-year-old Baltimore pupil Taki Allen told local outlet WMAR-2 News.

 

Baltimore County Police Department said their officers "responded appropriately and proportionally based on the information provided at the time".

 

It said the AI alert was sent to human reviewers who found no threat - but the principal missed this and contacted the school's safety team, who ultimately called the police.

 

But the incident has prompted calls by some for the schools' procedures around the use of such technology to be reviewed.

 

Mr Allen told local news he had finished a bag of Doritos after football practice, and put the empty packet in his pocket.

 

He said 20 minutes later, armed police arrived.

 

"He told me to get on my knees, arrested me and put me in cuffs," he said.

 

Baltimore County Police Department told BBC News Mr Allen was handcuffed but not arrested.

 

"The incident was safely resolved after it was determined there was no threat," they said in a statement...

 

[mavhc]

On 26/10/2025 at 08:16, sigma said:

https://www.bbc.co.uk/mediacentre/2025/new-ebu-research-ai-assistants-news-content

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory.

wonder what percentage of news articles themselves "had at least one significant issue"

[Ditto]

I've been using ChatGPT recently looking at legal case histories - I wish I didn't need to but heyhoe! It effectively quoted a case scenario that staggeringly matched a scenario I am reserching. I asked ChatGPT to provide the source of the information. The reply was "That example was an illustrative summary I drew from the mix of practitioner commentary and unreported/settled determinations you often see in legal briefings (i.e. the sort of unreported tribunal decisions or settlement outcomes that advisers cite when counselling clients). I should have made that clearer — apologies."

 

Some would argue that fits the bill of "You made it up!". I've got a lot of useful information and often with verified sources which I can check out, but the 'illustrative summary' isn't something I find engenders trust particularly! I suppose it apoligised and I believe I could ask to to state clearly when it is building 'illustrative summaries' in future - and it will remember. All part of the AI journey I guess!

[Ditto]

Oh dear. I asked ChatGPT to explain how it came to state a number of points on the topic alluded to above. It ripped itself to bits! I'm coming to the conclusion it needs to be told 'No inference, just referenced based facts". I can talk to so many people about something they know very little about or are perhaps Daily Snail educated readers, and let them talk (you know the word) but let's say infered deductions. I really expert better from ChatGPT, but perhaps that is unreasonable?!

[6Foot2]

Link: Will quantum be bigger than AI? - BBC News

 

Quote

There's an old adage among tech journalists like me - you can either explain quantum accurately, or in a way that people understand, but you can't do both.

 

That's because quantum mechanics - a strange and partly theoretical branch of physics - is a fiendishly difficult concept to get your head around.

 

It involves tiny particles behaving in weird ways. And this odd activity has opened up the potential of a whole new world of scientific super power.

 

Its mind-boggling complexity is probably a factor in why quantum has ended up with a lower profile than tech's current rockstar - artificial intelligence (AI).

 

This is despite a steady stream of recent big quantum announcements from tech giants like Microsoft and Google among others.

 

Broadly speaking, we tend to think about quantum more commonly in the form of hardware like sensors and computers, while AI is more software-based – it requires hardware to operate.

 

Put them together, and we might one day have a new form of technology that's more powerful than anything we have ever created… although the word "might" is doing some heavy-lifting in that particular prediction, warns Brian Hopkins, VP and principal analyst in emerging tech at research firm Forresters...

 

[mavhc]

16 hours ago, 6Foot2 said:

Link: Will quantum be bigger than AI? - BBC News

 

 

Not for a while: https://eprint.iacr.org/2025/1237.pdf

[Bromcom_John]

Apparently there are a few names in the sector who are surprised that quantum is still managing its legs as well as it is and were expecting it to have fizzled by now.
Claims of delivery at scale from at least one company are quite, interesting. Millions of replicable q-bits still sounds like something from a badly written Trek episode.
Phrases like "room-temperature photonics" just scream holodeck shenanigans episode

Edited November 7, 2025 by Bromcom_John

[psydii]

VR, AI, Quantum Computing, Fusion. It all comes around again and again.

 

This iteration of AI is just very good autocomplete, and has expanded its capabilities into non-text-based domains. 

 

Not to belittle the current AI, but basically they are just the neural nets of yore, with more maths that has only become viable to run thanks to modern GPUs. (and another two decades worth of comp-sci experience of playing around with them since they were last "the next big thing") I mean, can you imagine life with out auto-correct in Word? Thus it is becoming with the current AI stuff - but fundamentally its just predictions - not understanding. Also people seem to have forgotten about the inescapable effects once popularised as "chaos theory", and the reason that  the "context window" and the very fundamentals of GPU architecture together ultimately doom the progress of the current implementation.

 

Pretty sure we wont see  the next breakthrough* until someone releases a paper on using Quantum Computing to build, train and run the next generation of models, that is practical at scale.

 

 

 

 

[mavhc]

Not to belittle the current humans, but basically they are just the apes of yore, with more braincells that have only become viable to run thanks to walking upright. They can say words out loud buy do they have understanding, or are they just chemical reactions and electricity? They're just atoms bouncing around in space, they're not really alive

[6Foot2]

Very reductive.