AI weapons is inevitable and will be rapidly developed as we already see.
1 Like
I guess soon drone operated house breakins and crypto ransoms are not that far in the future.
1 Like
The only saving grace is those chips have become bloody expensive!
Not just expensive but simply bloody
Today special applications like drones or even tanks are used. But in a near future I think the deadliest of all weapons may be stolen by an enemy and used against us: cars.
Cars kill over one million people every year, even without being hacked. They could easy erase the population of a whole enemy country!
Cars can’t refuel themselves.
Neither can cruise missiles.
1 Like
Getting there
13 posts were split to a new topic: AI and Science Finction
This is an interesting take on quantum computing and AI: Quantum Computing: The Great Scientific Illusion.
1 Like
I spent 2 hours with AI helping me fix my taxes. I messed up payments a few years ago and recent years were amended and so I got updated statements for several years from the tax authority.
I scanned them all in, and the AI helped put everything in a table, worked out which years were under- and over- paid. Followed my instructions on how to re-allocate payments and prepared a table of the changes required and helped me to draft a letter to the tax authorities in German.
The whole thing took over an hour and would have taken twice as long without it. It also caught an error I made.
1 Like
You are using AI as a tool where you check its output - fantastic. The dangers are in AI generated “answers” where the user has no clue as to the reliability of the output. Obviously this happened with computers before AI, but recent AI advances have magnified the problems. These are inherent. If the AI learns from social media posts it will reproduce the prejudices displayed in those posts. If it uses intellectual property, it probably has not paid for that right and it has only poor ways to evaluate the quality of the data. Fortunately the medical profession seems to be getting to grips with the issues. Lay people using AI for medical advice are playing with fire. Many companies probably have no idea what they have signed up for.
2 Likes
I’ve used AI for medical advice and it has given me dangerously wrong answers (luckily I went to the doctor the same day!).
1 Like
You are far more trusting with your personal data and AI than I am!
1 Like
That was also my thought when I read his message, but perhaps he anonomised his tax data before feeding it in.
The tax scans are already stored on google drive, I send the email via GMail. The data is already accessible by Google.
Is there really a big risk if an LLM provider sees my tax account statements?
Had ine today, I needed some help in filling out a Spesen form.
Called the main office thinking I would be soeaking to a person.
No, I got an AI agent, a really good one and it took a few seconds for the uncanny valley to kick in.
Everything worked but our system is so byzantine snd I knew I had to klick some button but not which button.
It went through the usual. "Have you switched of snd on again, I know, lets clear the cache, yadda yadda yadda and after ten minutes of fruitlessly not getting anywhere I hung up.
Half an hour later I tried again, only this time I demanded to speak to a human, it the nd me that it understands but first we need to do…
Yadda yadda yadda.
I know now that telling an AI agent to go fuck itsself is not a valid command and that HR is monitoring.
Now we have to wear colorful and fancy shirts in order to go undetected.
https://www.dezeen.com/2026/08/25/digital-camouflage-shirt-ai-video-surveillance-simon-weckert/
2 Likes