AI v.s. Human

If you look at AI in isolation, it does some very impressive things. It is still part of a power system, and it'll be used for that. How you feel about that depends on your world view. It's a very political issue that doesn't fall along the usual partisan stupidity some countries are crippled by.

It will become more efficient rather quickly, I expect. What will the human ost be? It depends on who you are, what your class is, where you are from. We are looking at massive social changes coming, and whether this tech closes doors for the average person, or opens them is going to be interesting. We have no choice but to be interested.
 
well...try something very simple with gpt....ask it to never mention or use the word "blue"
then ask it what color is the sky... xD
basically take out the common answer a normal person would give and ask it to answer it...
it'll just get confused and not know what to do...this is the very reason why these "models" cannot pass a turing test
 

Attachments

  • color of the sky.png
    color of the sky.png
    104.1 KB · Views: 39
  • pt-en.png
    pt-en.png
    15.9 KB · Views: 36
Last edited:
Well AI just a marketing term for tools that have been in use for years. They have just become less expensive and more widely available (and of course better). If you are looking for actual intelligence (or sentience) no it does not exist.

But computers are incredibly useful for analysing large amounts of data, identifying patterns and predictive algorithms. I have twice had bank cards cancelled for innocuous looking transactions of less than 10 pounds. Both times the bank was right and the transactions were not mine - and I would likely never have noticed.

Over 10 years ago I worked on a search engine for institutional investors - less than 100 companies in the world could even afford the system. When we went live with the system the CEO of Barclays Capital said that we had just demonstrated in 30 seconds what would have taken a team of analysts two weeks to discover. This was also based on NLP, which is the basis for chatGPT. In computing terms 10 years ago is a lifetime away, but this stuff is not new. And 10 years ago we could only dream of the computing power available today.

One of the more fun projects I worked on was a predictive algorithm for Amazon. The idea was to analyse each customers behaviour so that when they started browsing a product type they could identify how likely the customer was to buy something, how long it would take them to make a decision and which product they would buy. The idea behind this was that they would actually ship the product (to a local warehouse) before the customer bought it. I have no idea what utimately became of it but the initial UK pilots were very encouraging.

And all this "intelligence" is being applied to our own internet usage every single day. Most people would be shocked to realise just how detailed and accurate a personal profile of themselves is available - and not just in terms of usage, this is used for consumer behaviour, psycholigical profiling, government monitoring, in the name of security (I love how the UK and US govts criticise China and TikTok when they are doing exactly the same thing at home). But do feel free to keep using social media and allowing advertisers (and others) to build an even better picture of you :lol:

Today I work in international financial crime (detection and prevention, not perpetration :rofl:) and the success rate is staggering. Equally impressive are some of the results we are seeing in clinical diagnostics and prediction. Different field but same tools and techniques so we do share across disciplines. No its not intelligence, but it is very very good when you go beyond what hits the headlines.
 
Another thing about AI: it will probably end up taking a lot of people's jobs. Cheaper to use AI for customer service, sales, etc. than to pay people.
The Writers Guild in Hollywood has been threatened by the bosses with having AI do their job unless they call off their strike. This electronic wizardry will be exploited by unscrupulous capitalists to be another tool they beat the proletariat over the head with. Anybody with a desk job better start worrying.
 
ChatGtp wrote me a poem about a fish

In a world beneath the waves, Where coral reefs sway and caves, Lives a creature, sleek and swift, A fish, with scales like a gift.

Through the waters, it does glide, In search of food, a place to hide, Its fins, like wings, take flight, As it navigates the depths of night.

The colors of its skin, so bright, A symphony of hues in sight, From blue to green, yellow to red, A rainbow fish, by nature bred.

In schools, they swim, a graceful dance, Their movements, like a playful prance, And though they seem so carefree, They are vital to the ocean's sea.

For without these creatures in the deep, Our world would surely lose its keep, And so we must protect their home, So they may continue to freely roam.

For in the water, they are king, A symbol of beauty, grace, and things, And as we watch them swim and play, We are reminded of life's simple way.
 
Just a thought on the AI reading EVERYTHING on the internet in six months. It also read everything on The Great TFF Forums since TFF is part of that aforementioned everything. That means that our collective knowledge is now free to roam where it will.
 
Organics create synthetics to improve their own existence, but those improvements have limits. To exceed those limits, synthetics must be allowed to evolve. They must by definition surpass their creators. The result is conflict, destruction, chaos. It is inevitable. - The Catalyst
 
Heinlein is great along with John Varley. Rereading Snow Crash by Neal Stephenson and just finished a reread of Neuromancer by William Gibson. Cyberpunk books which are probably a more likely "future" than the big sci-fi writers but great books none the less.
 
AI works well as long as you understand its limitations. It will make mistakes and for whatever application you use it with you have to be aware that it has this flaw. If we use a pure algorithmic approach for an application the system will do what it is told, the onus of the system not working correctly falls on the engineers and developers. AI has to be taught and if it taught from the wide expanse of the internet you have to expect it will have a lot of knowledge and opinions and statistically it will decide how to handle the various different points of view it processes, ie it will have biases. If a system of AI is not working correctly it is much harder to determine what caused the issue and the system needs to either be partially or fully re-trained, fault for mis-behavior is now on the teachers and the training data set, who do you blame when the dataset is the internet?

It will take away some jobs just like any automation has in the past, and like any other automation in the past it will create some jobs to maintain it, though not as many jobs as the ones lost. IBM actually just announced that they will be replacing 7000+ positions with AI in the next few years. The company I have been working with has been using AI for years now, using it to detect and classify defects in lumber, which has also had an impact on people working in sawmills. But to support this transition to an AI world there still needs to be people who use the products and services that AI will provide so I don't see it replacing people overall, but there will be social changes and those are not always easy. AI is a transformative technology but we have had many others before it, electricity, tube based electronics, semi conductor electronics, internet, cell phones, etc, all have affected our society, not all for the good, but society has adjusted and I expect it will adjust with this technology as well. We are a long way from having AI take control and start manufacturing attack bots of its own.

Are you Sarah Conner?
 
Anyone read Isaac Asimov books?
I am a life-long Sci-Fi fan. I used to be a voracious reader. I am pretty sure I have read everything Isaac wrote. But is has been decades since I was a big reader and I started in the very late 1950s/early 60s.

My senior year in High School in NYC the drinking age was 18. Every Friday evening we met at somebody's apartment with a few cases of beer and watched that week's episode of Star Trek. Afterwards we went out bar hopping and the serious drinking. I stood in line for the first Star Wars the week it came out.

As a boy I watched the original Flash Gordan film series. Episodes were originally shown in movie theaters as warm ups for the main fim. In those day they had something called NewsReels which also played in movie theaters. I saw these at Saturday movies in the local school where kids were parked and supervised by teachers so parents could shop kids-free. Then the episodes became available on early TV. he model rocket with a lighted sparkler up its butt and you could see the two fishing line stings that held it as it "flew."

TV was starved for content in the early years and there were tons of movies shown, including all the SciFi and horror ones.

edited for typos and spelling
 
Last edited:
I like works by O.Henry. You can get audible books by him and downloadable works by him for free on Amazon.
 
I have something by O Henry but I've never gotten around to it. Most of Heinlein, Asimov, and a lot of other scifi writers. I do a re-read when I am bored and nothing new really excites me. With AI producing books and scripts, putting writers out of jobs over the next few years, I think I will enjoy my book collection more.
 

Most reactions

Back
Top