AI - technical and ethical aspects

bikegremlin

To say it straight away: I'm a pessimist and fear that AI will mostly make a change for the worse.

It will definitely make huge changes, and make them surprisingly quickly. At the time of writing, an average Joe doesn't know what ChatGPT is. I expect that by the end of this year, it will replace Google to a great degree, unless Google makes something nearly as good (or "good", depending on how you're looking at it).

Of course, I could be wrong, and of course you can use it to help you and to make good stuff.

I wrote an article about my current experience and how I'm using AI at the moment, along with my moral dilemmas:
https://io.bikegremlin.com/34112/my-ai-experiment/

What are your thoughts on this?

Relja StillAnIdiot Novović

Chievo

AI would kill many professions soon (less than 6 years from now)

bikegremlin

@Chievo said:
AI would kill many professions soon (less than 6 years from now)

You are an optimist.
Some professions, like translators, are already being decimated.

Chievo

@bikegremlin said:

@Chievo said:
AI would kill many professions soon (less than 6 years from now)

You are an optimist.
Some professions, like translators, are already being decimated.

There is an AI in the dentistry which is making even easier everything so basically in 5-6 years things would change radically killing some professionals. I have no idea about another sectors i am speaking about the mine but i guess same would happen with many many sectors

lowendspiritxdax

While AI does indeed kill many professions, it is of great benefit to the majority of people. In the end it will accelerate human evolution, without doubt.

You can compare it to
1. Spread of education/literacy in the last few hundred years: church related people lost their job, as did kings and noble families, but farmers and butchers could send their children to school as knowledge isn't restricted to the nobles anymore. This accelerated science significantly.
2. Spread of machines and robots over the last 100 years, steam-engines and alike: horses lost their job, many workers in the clothing, food processing, farming industry lost their jobs to mass manufacture. But the whole world got interconnected with railway, boats and airplanes. Now you just have to press one button at home and you have coffee. You have cheap electricity and light instead of having to go out to the forest and collect firewood.
3. Spread of internet in the last 20 years: mail delivery workers and information brokers lost their job, massive amounts of local shops are loosing their business to Amazon, eBay etc.. But now everybody has access to pretty much all data in the world. Anyone can learn anything (youtube, udemy,...). Anyone can (in theory) do anything he wants and is not forced to go to the field to plant potatoes.

TL;DR with every new technology there comes benefits and disadvantages.
Some jobs will definitely cease to exist, while many new opportunities will be opened for the next generation of kids.
(e.g. 20 years ago there was basically no hosting business)

"in the past, everything was better"

wankel

@lowendspiritxdax said:
Some jobs will definitely cease to exist, while many new opportunities will be opened for the next generation of kids.

Thanks for the optimistic view :-)

The changes in the past did not (worldwide) concentrate and restrict new opportunities to the happy few introducing the opportunities.

There are only a few companies owning the current crop of AI, all fervently working to have their AI to the point where it can improve itself by itself. As byproduct they sell those services that replace current experts in several fields.

Replacing experts in itself is not a bad thing: now they can focus their intelligence and creative energy on new projects. As long as they have something to eat etc.

As long as "work done by AI" is taxable in the same way as "work done by experts", and revenue is not a one way stream to profit for a few worldwide monopolies but becomes redistributable wealth, there is room for a brighter future.

You may be familiar with "Manna" by Marshall Brain for the distinction. If the eutopian trouser of time has your interest, I can wholeheartedly recommend the Culture-series by late Iain Banks.

terrorgen

@wankel said: There are only a few companies owning the current crop of AI

Welp, we are back to "capitalists owning the means of production" again.
Karl predicted this.

XXSL

The professionals who are easily killed by AI are not real professional.

Chievo

@XXSL said:
The professionals who are easily killed by AI are not real professional.

It depends. For the moment in dentistry i could save 5 min gross per some special cases. In big dental practices 5min per patient could kill a dentist jobs or reduce his hours . Is this a professional? Yes! If the machine AI is not doing his work .he could be working full time instead of part time or not working at all. I guess that in every profession depends but more or less true for many of them.

techdragon

The only thing left is to agree on The Blueprint for an AI Bill of Rights.

havoc

Mixture of three things:

1) Personally excited - it's cool tech.
2) Think it can be the next leap forward for humanity like electricity was (think lets chatbots and more this )
3) Worried the social system isn't ready. i.e. what are we going to do with all the people that can't compete anymore. Not all countries can afford UBI

I've decided this is happening one way or the other & may as well roll with it. My employer is rolling out loads of AI, building AI stuff in spare time and tilting investment portfolio towards it too. Also used AI today to build a clustering algo to analyze a couple thousand mails at work...which definitely felt like magic.

vyas

Too many AI related changes at workplace

Mostly positive so far. Productivity, creativity, learning.

Might post a longer "How AI helped me pivot my business" post later.

vyas

Not everything related to AI needs to be very hi-tech or complex. Sometimes, a quick coversation over a coffee/ beer can help form new business relationships or personal connects. Enter AI to help, Atleast for now.

Below is the output of a five minute conversation with a neighbour. He runs a stock trading group, and wanted an image for Instagram. He was okay with base 500x500 px resolution image. (Seasoned "offline mode hand" new to IG... yes, such people do exist even in 2024)
I asked him a few qualifier questions (style, format, scale, theme, etc..) put it into AI imaging tool, sent him 4 options, 3 of which he liked, I scaled them up to 2000x2000 (AI upscaler), send off.. all in 5 minutes!

• Image gen, 1:1 messaging via phone;

• Scaling up / sending via CURL + API

(Not showing off- a week ago I had NO idea that was possible. Now I do :-) )

Below are the images Links will be disabled in a week's time if not sooner.

https://img.gaatha.me/i/3GMJeC
https://img.gaatha.me/i/pgSd69
https://img.gaatha.me/i/c1Kp3j

---

Apologies for a second post. Late by a few minutes in editing the one above.

bikegremlin

@vyas said:
Too many AI related changes at workplace

Mostly positive so far. Productivity, creativity, learning.

Might post a longer "How AI helped me pivot my business" post later.

There's a story from our Nobel prize winning novel "Bridge On The Drina River" (by Ivo Andrić):
It takes place in Bosnia, during the Austria-Hungarian rule.

A guy from a remote village is thrilled with a new railroad that Austrians built. "I can get to Sarajevo in only two hours now!"
His friend says: "Do you think Austrians built that railroad so that you could get to Sarajevo more quickly?!"

The benefits that you (and I) can get from AI are needed in order to not get any rebellion against it.

Also, the AI development looks a lot like the nuclear arms race - everyone is trying to get ahead, and control is not a priority.

So, such a powerful tool can easily be abused to enslave and control people - either deliberately or by good intentions.

At least that's how I see it. Would love to be wrong about it.

Diwakar

Whether AI replaces Google or not is difficult to predict, but it's true that AI is rapidly evolving and making significant strides. It's crucial to approach this development with both optimism and caution, ensuring its development and use are ethical and responsible.

I appreciate you sharing your article. It's valuable to see personal experiences and perspectives on this complex topic.

lowendspiritxdax

@bikegremlin said:
So, such a powerful tool can easily be abused to enslave and control people - either deliberately or by good intentions.

I am worried with the world leaders we currently have (every major country haha)

SenseiSteve

AI is so new to me that I soak in discussions such as this as much as possible. I have experimented with ChatGPT and found it useful for updating articles, but my real interest is in graphics. I see both the good and bad attributes to how AI is implemented by governments. For sure, AI is not going away and will continue to grow by leaps and bounds, so much so that you need to jump on the bandwagon now or risk being left in the dusk. Clearly, Google has already shown its bias in the AI arena, as I suspect other social media platforms. Seems odd to say, but I'm both optimistic and pessimistic about where AI is taking us.

crunchbits

@bikegremlin said:

@vyas said:
Too many AI related changes at workplace

Mostly positive so far. Productivity, creativity, learning.

Might post a longer "How AI helped me pivot my business" post later.

There's a story from our Nobel prize winning novel "Bridge On The Drina River" (by Ivo Andrić):
It takes place in Bosnia, during the Austria-Hungarian rule.

A guy from a remote village is thrilled with a new railroad that Austrians built. "I can get to Sarajevo in only two hours now!"
His friend says: "Do you think Austrians built that railroad so that you could get to Sarajevo more quickly?!"

The benefits that you (and I) can get from AI are needed in order to not get any rebellion against it.

Also, the AI development looks a lot like the nuclear arms race - everyone is trying to get ahead, and control is not a priority.

So, such a powerful tool can easily be abused to enslave and control people - either deliberately or by good intentions.

At least that's how I see it. Would love to be wrong about it.

Great example. I know I'll be viewed as biased (pro-AI only) due to slinging my wares, and there are some aspects of it that I do like. Of course, I was slinging it before "AI" and thought there are many use-cases for the raw infrastructure before the current year "AI"-craze. There will be a reckoning at some point with the build-ups going on everywhere, but it's too hard for me to say what part of that curve we're at yet. Likely still too early.

That being said, most of the stuff isn't really... AI, or at least not what I would consider AI to be. It's just faster/better processing of datasets. Extremely personalized google. There are some ways in which I can see it being insanely productive and helpful without harming (some of which I've been exploring internally). For the others, well, did everyone die and all jobs disappear when the car replaced the horse? I think it's just likely you see a shift where the more 'menial' tasks go to. Street shit sweeper became Jiffylube employee. If you don't have experience in a data center dealing with these things I can tell you there will be a surge in entry-level tech positions to maintain these units. "Can you use a datavac?", "Do you know the difference between EPS and PCIe plug?", etc.

I still think things that might be "at risk" (say copywriting) will require a human editor/proof reader and they will just become 10x more productive. Price-per-word/article (however it is valued) will drop, you might cull some of the weaker in the field (i.e. part-time real estate agents in 2009) but I just don't see what we've got being Skynet-tier yet. It just feels like we went from hand-crank drill to first power drill.

terrorgen

Ethicists need to get ahead of the game though.
Lawmakers as well. And they in most countries don't even seem to understand tech.

Chievo

@bikegremlin said:
To say it straight away: I'm a pessimist and fear that AI will mostly make a change for the worse.

It will definitely make huge changes, and make them surprisingly quickly. At the time of writing, an average Joe doesn't know what ChatGPT is. I expect that by the end of this year, it will replace Google to a great degree, unless Google makes something nearly as good (or "good", depending on how you're looking at it).

Of course, I could be wrong, and of course you can use it to help you and to make good stuff.

I wrote an article about my current experience and how I'm using AI at the moment, along with my moral dilemmas:
https://io.bikegremlin.com/34112/my-ai-experiment/

What are your thoughts on this?

Relja StillAnIdiot Novović

Just taking some red bull with friends and we were speaking about this topic so basically i just discovered a new tool from the @bikegremlin area the balkans which is taking some steps . so basically helps dentists in order to find more caries more periodontal cases saving time and " possible making" of more money. For the moment it is "expensive" . I mean you pay a certain amount of money to diagnostic better with bettet accuracy but the patients may be/probably are not ready to accept such treatments YET . hopefully in the near future we would have more players on the market and the price would be "lower" or more friendly.price let.s say.

So more accuracy better than a human brain+ eyes.

Another way is sending sms or whatsapp or another automatic marketing automation could be done with it( AI) so less time doing burecreacy.

It works 24/7 another benefit and translates content well . another point

Ethically. Better diagnostics better treatments. But you need prepared patients as well in order to accept what the machine says . it depends from the perspective average age etc and vision about technology

wankel

@Chievo said: Better diagnostics better treatments. But you
need prepared patients as well in order to accept what the machine says

I read/heard about this weather forecast AI by Google (not in this thread, I hope, might be though...) that takes two snapshots of weather info six hours apart ("now" and "six hours ago"), and based on that, can give a better prediction for the next ten days than current "old fashioned" forecasting models. For a fraction of the computing cost.

With your dentistry website, would it be interesting to have people upload a smiling photo full of teeth and a description of their symptoms, connect it to AI for pattern matching, and then return a score for one or another dental problem and an urge (or not) to visit a real dentist? That way not the patient, but the dentist has to accept (or verify) what the machine says.

Chievo

Hi ! Yeah it would be interesting option not sure which ai for pattern matching could be used for that mission. Please,name at least one tool for ai pattern matching in order to see how it works and if could be applied . thanks for the feedback.
For the moment the dental ai (which i was speakin) is just for radiographs but true not sure which is their AI Based on . Thanks once again !

havoc

Well if I were a programmer I'd officially be shtting bricks as of right about now. Cognition just showed off a tool that can do some fairly complicated stuff autonomously. Not generate code...but like write it, research errors, look up documentation, figure out APIs, deploy it, write documentation.

ahnlak

@havoc said:
Well if I were a programmer I'd officially be shtting bricks as of right about now. Cognition just showed off a tool that can do some fairly complicated stuff autonomously. Not generate code...but like write it, research errors, look up documentation, figure out APIs, deploy it, write documentation.

Given that more and more programming-related questions I now see start with "I have this code that {random AI shitpile} wrote for me...", I can only see this crap generating more work for me, fixing the messes it makes.

I don't see programmers going the way of wheelwrights any time soon.

havoc

@ahnlak said:

I don't see programmers going the way of wheelwrights any time soon.

The programming gigs involving real heavy lifting are presumably safe for quite a while still. Think the risk is more that some get pushed out of the boilerplate roles and thus there is more competition for the remaining roles?

Regardless, interesting times ahead.

ahnlak

Meh; the boilerplate generated is mostly junk - at some point in the next couple of years there's gonna be some catastrophic system failures somewhere, and it's going to turn out to be because of some generated code crap that nobody checked/understood, and someone will get sued hard enough to make everyone think very carefully about using it.

Luckily, I'm ancient enough that if I'm wrong I don't care

Neoon

@havoc said:
Well if I were a programmer I'd officially be shtting bricks as of right about now. Cognition just showed off a tool that can do some fairly complicated stuff autonomously. Not generate code...but like write it, research errors, look up documentation, figure out APIs, deploy it, write documentation.

Nah, 13% and I remind you, Lies, damned lies, and statistics.
I wanna know how they benchmarked it, prob rigged like Intel and all other companies do.

LLM's can't solve big problems yet write entire projects on its own.
You can use it to fix small stuff or write some stuff but that's it.