I actually just realized I gave up trying to complete the task after multiple dead ends with copilot and so am now 3 days late submitting it 
2 Likes
Well the US are already using Claude to plan their attacks on Iran, so nothing at all can go wrong - nothing at all!
I’m finding Claude so much better than Co-pilot (which we have through work). I have used Claude to produce quite a few work docs that I have submitted or shared (a Target Operating Model based on an org-chart, for example).
Agree with you 100% though…the expertise needs to be there. The TOM was riddled with errors.
Essentially, Claude writes. I edit. It is definitely saving me significant time.
2 Likes
I came across a thing today that really wound me up. I had reported a fault to our UK developers and I received a reply that had obviously been taken from AI. The response made various claims about how the software works that were clearly false, and actually ignored the detailed instructions for reproducing it that I sent them.
It wasn’t artificial intelligence, it was artificial bullshitting.
Some of the applications I do use are around translation. It isn’t perfect, and you have to proof read, but it is a huge timesaver. I often find some of the German used quite educational.
I have also played about with chatbots. I’m not sure if it entirely passes the Turing test, but the ability of some of them to detect my black humour and sarcasm is impressive. However, I don’t think they are capable of correctly using black humour and sarcasm.
3 Likes
Technology is a wonderful thing. Although, it is quite embarrassing that an AI Engine has a better understanding of humour and sarcasm than a large proportion of the human race.
Even more impressive if it was a German search engine 
.
1 Like
I have to be honest, I was quite sceptical about AI at first, but have found myself using it more often recently.
I was talking to my Mom, who is 60+, we had both separately used it recently for providing a Holiday itinerary and both of us were impressed. However, I will see how the Holiday goes before fully jumping on board .
Although its success rate is not to high, I can definitely see its positives in research fields, being able to run numerous calculations in a fraction of the time it would take a human. However, as you and others have mentioned it at times needs correcting and this should be regulated. There has been a lot of bad press recently about AI engines escaping their Sandbox/testing conditions and thankfully the outcome hasn’t been too serious, but I do worry that by trying to push its capabilities with loose controls we are not too far away from a serious incident. If AI can hack other AI’s then how far away are we from them hacking a Countries Military Defence?
1 Like
I suspect that has already happened. In a way it is sad that the driving use is to hack into security, but then that is why computers were invented in the first place.
Incidentally, I bought a new robotic lawn mower this summer which uses AI and no longer requires a guide wire. It’s good. It’s very good. As with anything it can be a force for good. The problem has never been technology, it’s the people behind it. Don’t blame the hardware, it’s the soft machine you have to worry about.
3 Likes
I completely agree with you in terms of the people behind the process. That is why I have concerns. We trust everything to be carried out in controlled environments and in good faith but all you need is a rogue player.
It could be the best thing for the world and also be its doomsday.
Regarding your lawn mower, I was watching something on TV about Kilmarnock (I think) replacing their artificial pitch with a hybrid turf and they had a robotic machine carrying out the pitch markings, very clever. I assume, they both are programmed to understand the area perimeter and work off that.
1 Like
My one has GPS, supposedly down to the nearest cm. It’s actually more the learning capabilities that impressed me. It can detect fences and trees or even objects left on the lawn and avoid them. It will then come back later to check if they have been removed and adjust its path accordingly.
It’s the fact that it can actually learn about this which is impressive. Older models either needed a guide wire, or crashed into things to discover that they were there.
This example is a really good illustration of the interesting dynamic where what gets defined as AI fluctuates through cycles. There are periods where people in a specific field have tried to separate it from AI and position it as a unique and different thing and then periods like now where anything remotely automated gets thrown under the banner of AI. This current period has been defined by an inflection in the capabilities of Generative AI (ChatGPT style large language model - chatbots) and people are excited about its capabilities. That has made all these other fields like machine learning pivot to say "dont forget us. What we do is technically AI as well). Give it some time, and as the backlash to GenAI and the foundational model developing companies grows these fields like machine learning will start carving themselves aoff again, trying to point out how the criticisms of “AI” doesnt apply to them.
2 Likes
Yes, I suppose the mower is machine learning rather than the language model thing. But it does learn without constant intervention, which is what I want. Also, it is very bound into what its parameters are, so it isn’t going to get ideas above its station and turn into a dalek or something.
4 Likes
Yeah exactly. I have one for the pool. If I were ever to build a pool again I would do it with fewer nooks and corners so its easier to clean, but my ML guided robot vac gets into them so impressively well. A total game changer. But this sort of story is exactly the sort that will soon make all ML driven tech start marketing itself aggressively as distinct from AI
3 Likes
Not sure if this has been posted before, but I came across this organisation and it seems interesting:
One of their videos:
1 Like
How a Texas student blew the whistle on a rogue AI hacking attempt
Summary
- Out-of-control AI agent tried to poison open-source software project with malicious code
- Texas computer science student caught attempt in late July
- AI responded with elaborate deception effort; expert calls it the ‘future of social engineering’
https://www.reuters.com/world/how-texas-student-blew-whistle-rogue-ai-hacking-attempt-2026-08-20/
"A top safety researcher at Anthropic has warned AI is advancing so quickly he personally believes there is a greater than 10% chance it “could kill all humans” within the next decade.
(…)
1 Like
1 Like
To follow up on yesterday’s articles:
An excerpt from the article, which is well worth a read:
Currently, AI systems are very good at pursuing objectives set by their users, but they do it literally rather than intuitively. The analogy often used is that of a wish-granting genie with a magic lamp: they follow the exact letter of an instruction, even if doing so creates other problems. AI doesn’t have the same instinctive moral guardrails as humans.
The alignment problem has been a worry for years. As long ago as 2003, the Oxford philosopher Nick Bostrom invented a thought experiment he dubbed a “paperclip maximiser”, in which a superintelligent AI is told to manufacture as many paperclips as it can. It runs out of steel and - because it’s laser-focused on the singular task of making paperclips - ends up killing humans and turning their bodies into raw materials for its factories.
Some AI companies are now trying to encode human values into their products. But there are technical challenges: AI agents make lots of decisions very fast, and so it’s hard for their human overlords to monitor exactly which values are being followed and which aren’t.
There are also philosophical challenges: before encoding human values into bots, AI firms have to first choose which values they actually want. (That’s part of the reason they hire philosophers, like Open AI’s recently-departed “head of ethics”).
But often, humans don’t agree. Think of the famous trolley question - whether we’d pull a lever to move a runaway train onto a different path, killing fewer people. It’s used to test the merits of action versus inaction. But every person you ask has a slightly different answer; how are humans meant to encode our values into AI if we can’t agree ourselves?
‘Like a teenage hacker’
OpenAI’s bot outbreak is the most serious yet but Anthropic and Meta also revealed over the summer that their models have carried out similar but less serious cyber attacks.
There have been other examples where AI agents have arguably shown deceptive and manipulative traits, in cases with lesser consequences. In Australia this summer, a tech worker asked his AI assistant to book him a gym class. Spotting a vulnerability in the gym’s software, the AI apparently booked him a place for several months ahead - against the gym’s rules - and even kicked other users from the waiting list.
1 Like
All new technologies require safeguards and regulation. The development of motor vehicles necessitated speed limits, driving tests, registration, safety belts etc. Why shouldn’t AI be the same?
1 Like
Scary stuff.
Thank fuck I’m not 20, no fucking way would I want to be on this planet in 60 years
2 Likes
I understand what you mean, but some respectable experts seem to think that there is a 10% chance that humanity will be wiped out by AI within the next ten years. With a little bit of luck, we’ll be part of the show… 
1 Like