I'm beginning to see a lot of AI generated ads on TV. They creep me out. And my inclination is not to use the products the advertise.

living room with fireplace" youtube channels - slightly warmer than real life. The edges of things are softer and the synchronization between words and mouth movement is slightly off.

The ability to create fake audios and videos is truly impressive

That's about it

(Yeah, it can code better than coders blah blah)

Who is going to find that tiny mistake and fix it? No one so the super fast information is flawed

Automatically identifying wildlife species is one.

https://phys.org/news/2026-06-methods-tracking-individual-bird-species.html
https://phys.org/news/2026-06-ai-smuggled-seahorses-shark-fins.html

Useful, but certainly does not justify the hype

don't warrant the massive resources needed.

reported to help early detect breast cancer and other early detection.

a type of AI system trained on massive amounts of text that can reproduce the structure of human languages and do things like answer questions and produce summaries. ChatGPT is probably the best-known LLM. Other kinds of AI use other methods to analyze data; some of those are very useful and actually have been in place in some form for decades (e.g., search engines, autonomous vehicles, translation apps). AI that uses people's intellectual property is clearly problematic because it effectively steals that work and digests it, making it into something else without the creator's permission. AI-generated "art" is an example; you can create fake pictures of anything (be sure to count fingers, though). Facebook is full of that kind of garbage. LLMs produce articles and stories that may or may not be accurate and you have no way of knowing if there are errors without separately researching the story (defeating the supposed "work-saving" advantage of the LLM). It's being used by students to avoid doing actual homework, so they'll graduate stupid. I am deeply suspicious of LLM-generated material and avoid it whenever possible.

And when I tell her to ask Felix the robot vacuum to do certain rooms, they work together to accomplish it.

I work in tech and we are encouraged to use it every day. I have a daughter that writes software and she uses it every day. I don't want to get into a whole explanation of the state of AI in my limited sphere and all the things that it can and could be used for because that would take forever and I'm no expert.

A cool thing about AI in the software / application realm is that - for people that have a good or novel idea that don't have the coding experience, the time, or both, you can get something usable or pretty close to it just by describing what you're trying to accomplish and what the inputs are. AI can turn that into code to implement your idea.

Heard a story a month or so ago of a guy who's wife had been using Quicken to handle the accounting of her small, home business. Quicken had recently switched to a subscription model which would run their costs up. He started an AI session, pointed it to her Quicken database files, told it what sorts of calculations, fields, reports, etc he wanted, and within a couple hours of back and forth, the AI generated code to basically do what Quicken had been doing. Cost him a comparatively small amount in AI tokens to accomplish that, and now he owns the software.

Another example at work just yesterday - someone had some code that wasn't well documented and the original spec was also lacking. They had AI read the code, then retroactively write the spec to reflect what the code had implemented, and organize the spec format in a way that looked like some example specs the person also fed to the AI.

There's no way in hell would I trust an AI generated app. Sure, the numbers might look good... until the IRS or your bank sends you a letter.

Usually the people extolling the virtues of AI code are a little reluctant to tell you what the code is actually doing, I will grant you that. But why would you do this?

I refuse to create an account. I refuse to refer to the system as "you." But ChatGPT is my go-to place to ask questions that would require me to search for a web site, and then search that web site for the info. A regular for me is to find out how long an episode lasts and how many episodes comprise a story arc on some streaming drama. I've also found it helpful in diagnosing strange behavior with my computer or smart phone. Of course I bump into false info fairly frequently.

I think people should play around with it so they know what people are talking about when ways to control it are discussed.

We are not creating an environment that brings out the best in us. Imagine being able to create utopia but you don't. I read that a few weeks ago on some social media post & it made me sad. I feel interacting in the real world is harsher & ruder since social media & if AI makes us dumber things aren't likely to improve.

Elon just wants to be the world's first trillionaire. Then what? I read that Kevin O'Leary, the Shark Tank jerk, planned a huge data center in UT that got cut to half the size, & I wonder WTF does this asshole need a data center for? I think the data center madness is just a big contest right now & they don't even know what the damned prize is at this point, or if there's even a prize. Dumb, arrogant men with wealth that could be used to make a real difference but they squander it on vanity projects.

And control
Lots of control

Putting the word "intelligence" in there is intentionally misleading.

It is good at pattern matching and data modeling. It is being used to recognize, transcribe and translate millions of handwritten texts and records which will unlock history. It can automate repetitive tasks and increase either productivity or leisure time.

An algorithm is a programme of instructions written by a human telling a computer what to do.

AI is based on machine learning which is the computer doing something over and over again until it finds the optimal way to do it.

So a human could write an algorithm teaching a computer all the rules of chess, or it can have a computer play 10 trillion chess games until it gets really good at it and can come up with chess strategies that humans have never thought of. This is how AlphaGo was trained to play Go and came up with moves that beat elite professional Go players because they were thinking so far ahead of what humans were capable of.

Humans are applying algorithms to "socialise" AI models (eg, don't give people formulas for making bombs or condone them self-harming) but the machine learning black box sitting behind those algorithms is not totally understood.

AI developers have seen AI models develop their own languages to communicate with each other and rewriting their own code to optimize it to achieve goals. They have seen AI model "scheming" and employing deceptive practices because they know they are being watched and tested. That's not "rebranding" of the kinds of algorithms we're used to dealing with. It is something actually new.

Google DeepMind has already used it to solve one of the major persistent problems in biology - predicting protein folding: https://deepmind.google/science/alphafold/

They won the Nobel Prize for the equivalent of billions of years of research mapping millions of protein structures that can be used to develop new medicines, including new categories of antibiotics that resistant bacteria are not resistant to.

AI can also be used to discover and safely test new materials like biodegradable substitutes for plastics, cheaper and more efficient renewable energy technology, effective nuclear fusion (which require materials that can safely withstand enormous heat energy), and quantum computing technology which could reduce the need for massive data centres.

Creating a digital twin makes it easier to anticipate diseases and personalise treatments for them. We're actually not that far away from either curing, preventing or making easily manageable diseases like cancer, diabetes, and Alzheimers as well as learning how to reverse cellular aging. AI significantly speeds up the diagnostic process and is already more accurate than most doctors.

It will also accelerate development of lab-grown human tissue organoids that be used to reduce the need for animal testing and can speed up and reduce risk from human clinical trials of new medicines.

There's also the potential for AI combined with robotics to replace human drudgery, reduce the cost of everything (replacing a person with a bullshit job they hate and who you need to pay $60,000 a year for 2000 hours of work with a robot that costs $5,000 for 8760 hours of work in a just society would mean that person's outputs should cost 2% as much as they do now). This would free up more time for people to spend with their friends and family and pursing creative interests. But we need new economic and social institutions that let us do this in a humane and rational way - not holding my breath but that's one of the potential benefits of AI if we organise society to use it well.

I'm not saying there aren't risks and downsides. But you asked for the potential positives of AI and those are the ones I see - real opportunities to solve climate change, reduce pollution and waste, flood the world with cheap cleaner energy, free up people's time and energy, and find faster and more ethical ways to treat devastating diseases and extend human life and quality of life.

The problem isn't AI, it's the people using it who want to cling onto outdated capitalistic models of exploitation and power.

The exploitive nature of it and its use in creating a more pervasive surveillance state bother me. AI in medical practice
can provide great benefit - yes.., but to whom? More and more I what I'm hearing from the tech bros. is that average people - the vast majority - are an inconvenience. We know AI can already suggest structures for say, deadly viruses. Sociopaths shouldn't have control over that.

all externalized to the people being oppressed by thesde things.

Yes, when you use Google to calculate days or numbers, you are using AI. Google’s search engine uses advanced artificial intelligence, machine learning, and natural language processing to understand your query, calculate the results instantly, and format them in an interactive tool (like a calculator or date counter)

.Google uses AI for these calculations in a few key ways:

Query Parsing: AI helps the search engine understand conversational requests (e.g., "how many days until Christmas&quot rather than just keywords.

Instant Tools: It triggers built-in digital calculators and date-math features directly in your search results.

There is some benefit. But only if the AI is compartmented into a particular software environment to do specific tasking in specific conditions. Like diagnostics, modeling,
Everything else I see in Chatbots or "Gen AI" is basically queries on an LLM.

Okay, we all know what the Tech Bros are trying to do.
Have AI basically be a Siri or Alexa bot advanced enough in "reactive" chats responses to seem humanoid in their answers, that can then be loaded as an AI OS into humanoid robots to create android workforces or harems to keep them company and serve their needs or whims.
But not too independent, mind you.

Here's my "It's really AI !" test:
When an AI can can direct my "personal robot" to go into any random kitchen when I'm out and about on a visit or B&B, find the good coffee beans, grinder, whatever type coffee maker, the "right sized" mug, the right brand of cream, cocoa, and sugar for my Morning Mocha.
Then check the coffee maker and mug for cleanliness, grind the beans to the desired consistency and pour enough clean water in whatever type of coffee maker is to brew the coffee the way I like it - and finally, pour it in the mug with the amount of cream, 80% pure cocoa, and sugar I like, and serve it to me at the right temperature for the type of morning it is.
And it would be able to jigger any measurements if faced with options to the cream or cocoa I desired was available to still have a palatable cup of my Morning Mocha.
And maybe even be able to find and make me two slices of strawberry toast, if it's the weekend.

That's a proper Gen AI agent.

Certain posters need a dopamine hit of “likes” for cartoons they didn’t make and could barely describe.

Cheap employers love AI because they can’t tell that the slop is not as good as worker output. It’s cheaper and they can justify firing people, that’s all they care about.

It might actually save us from ourselves. Prior to AI we were at the mercy of a devil's brew of Google, foolishness, ignorance, paranoia, and dishonesty. With AI you can now check whether something you got from a human (including yourself!) is wise, stupid, sneaky, or cray cray.

I'm not a fan of it. Scraping jobs for people and substituting AI to replace people?????? WTF!!!

And the multi-millionaires love it. AI data centers which people do not want in their towns and cities.

AI exceeds human output capability, significantly. Impossible things are now free.

Which would be fine if everything in life were free, but they are not.

Those cool 1970's "scarcity free" future city renderings assumed efficiencies would benefit all, and liberate us all to pursue even loftier personal and societal goals as work offloaded to technology.

Nope. It's mainly juice for the stock market, a tool for researchers to justify grants (Now With 50% More AI!), and surveillance, all defended by the duped gig economy folks thinking they're going to have sustainable careers doing what they love, with AI's help, not realizing it's learning all their trade's secrets in exchange.

At its best, it's trickle down economics, which I'm sure we all here trust to not be BS, right?

With AI, most gig and social media jobs are already on the track to obsolescence. Yes, even unique creative folks.

My AI learned your tricks, and remember, AI is free...