Telling displaced workers to “just reinvent themselves” is an inspiring individual goal, but an impossible large scale solution when tech evolves faster than humans can retrain.
1 Like
I’ve been thinking quite a lot about this idea as I’ve read through this discussion. I think that we first need to agree on terminology and language before we can come to an understanding of what it is that we are talking about when we talk about “AI”.
Technically, AI is a field of study involving building systems that attempt to mimic human intelligence. I think this is a good and worthy pursuit, and that we should be spending time trying to make our machines smarter and better at the tasks we want them to do.
Machine Learning, as I understand it anyway, is a specific area of AI research dealing with teaching computers to predict future states or actions based on previous data. Great work has been done here in image editing, like the ability to remove a telephone pole from a picture that you didn’t mean to capture.
Large Language Models, the LLMs that I think most of us are actually talking about here, are a specific area of Machine Learning, (well, I guess Deep Learning, again under Machine Learning). LLMs are the ones that take massive amounts of data to train, and the training and hosting of the LLMs is what’s driving the data center boom across the US.
What we see in Claude, ChatGPT, and the like is the application of LLM technology to a product. I take great issue with how the LLMs are being developed and deployed by these companies. From Grok building methane gas turbines in Memphis to OpenAI’s supposed hack of Hugging Face. They are burning cash (and the environment) at an alarming rate, and, in my opinion, can’t be trusted.
I don’t believe that these companies are going to build a Artificial General Intelligence (AGI), but I do believe that they need as many people in the world as possible to think that they can so they can keep that influx of cash coming in. When they claim that AI can destroy the world, or that AI is too dangerous, what they are really saying is “look at how amazing our product is! Give us more cash.”
But, when you say “keep in mind the immense benefits of AI”, I don’t think you are really talking about the application of LLMs. I don’t think these chatbots are going to find a cure for cancer. But it is possible that another branch of AI research can. Other very specialized models that are trained on cancer and medical research may be of great benefit to mankind. The same could be said of other hard problems like climate change and world hunger. AI very well might be able to find solutions for some of these problems, but it’s probably not going to come from a company that’s more interested in cash flow than scientific research.
So, yea, I think that there very well may be immense benefits to “AI” in the future, but probably not from chatbots. I also think that we should be far more careful and deliberate with how we are applying these technologies than we have so far.
Well, we can hope.
4 Likes
This is an excellent sophisticated article. Well worth a read.
Speculation: Consultants were brought in because folks who were using the software systems that would crash would otherwise have had to be trained in computer programming. Which would have cost 10x as much to fix, at 10x times as long.
In case it is important, I have no supporting research for this speculation. But, to bring it to present day … Perhaps we should arm a cadre of consultants properly versed in ethical human behavior and (pre-emptively) task them to oversee the equivalently unversed AI leaders, thereby keeping them from (naively or maliciously) committing ethical abuses with their software (… and commensurately lowering the global anxiety by a factor of 10x).
–
JJW
https://www.nytimes.com/video/watch
Bill Gates leans into the dangers of AI in the hands of bad people rather than AI’s wandering off on their own. But he is convinced that AI is extremely dangerous if it is just allowed to develop without regulation.
I just ordered a Mac mini M6 so I can understand AI better.
I am very much talking about LLMs (as well as other forms of AI).
The benefits of LLMs are here today in any knowledge based discipline - medicine, law, engineering, academics, almost every other profession. I am particularly referring to the ability of LLM to locate primary-source information. While I am skeptical about AI replacing such workers - human judgment will always remain essential - AI can locate academic articles or caselaw infinitely more effectively than using a paper library or even compared with a non-sematic internet search.
AI may not literally “cure cancer” but if it makes it practical and possible for your doctor to identify credible published articles pertinent to your treatment then that is an immense step forward.
I’m sorry, but that is hugely questionable.
While there are certainly real benefits, they are heavily outweighed by the downsides: compounding strain on climate change, the delayed retirement of fossil-fuel plants, cost inflation across the entire electronics supply chain, and the downstream increase in the cost of living. The positives right now simply do not justify the negatives, and there is zero guarantee of when; or if; that balance will flip. If the timeline slips, consuming our remaining carbon budget for speculative digital tools risks pushing planetary tipping points past the threshold of human sustainability.
What has been striking is how quickly governments that spent years preaching Net Zero, demanding the phase-out of fossil fuels, and warning of catastrophic temperature rise have quietly ditched those standards the moment AI entered the picture. It is equally worrying that society foots the bill through higher utility rates and degraded resources while tech giants privatize the profits.
All this is significant to warrant caution, and much faster regulation and making AI companies pay the real costs.
Edit
Of course if AI fails to produce and stuffs up our environment, it may cause an unintended extinction event 
.
The benefits of AI are objectively unquestionable.
Your point is that the downsides may exceed the benefits. That is certainly worthy of discussion.
That said - eons of history demonstrate that regardless of the net risk/benefit analysis, society will proceed to adopt the technology.
Calling the benefits “objectively unquestionable” overlooks who actually captures those benefits versus who absorbs the costs.
More importantly, the idea that society blindly adopts every technology regardless of risk isn’t historically accurate. We routinely halt, restrict, or phase out technologies once the net harms become apparent—leaded petrol, CFCs, human reproductive cloning, and chemical weapons are obvious examples. Adoption isn’t an unalterable natural law; it’s a series of policy and societal choices.
2 Likes
Again you are [understandably and correctly] pointing out that there are risks as wel as benefits. I agree. The benefits remain unquestionable. The question is whether the downsides exceed those benefits.
Human reproductive cloning was recognized as such a huge risk that it was prohibied before it ever got started. Yes, there are some other examples in that regard.
Leaded fuel, CFC, and Chemical weapons were only regulated after (or because) alternatives appeared which were at least as effective. None of them have been banned outright - though all do have very significant restrictions.
At this point there is no equally effective alternative technology to AI.
We’re obviously not going to agree. Statements like ‘unquestionable’ lack a broader perspective as they’re not beneficial to everyone.
As we’re going round the houses, I’ll bow out of further responses.
2 Likes
That’s quite a list of what our society has lived through in some form or another. For me, I don’t want to live through any period of instability. I can’t say whether AI + climate change will be as bad or worse than your list, but my gut says it will be game-changing.
I think the US will become like Mexico, with 1% owning most assets, a small middle class of doctors, lawyers, accountants, merchants, etc., and 85+% being desperate. I love the Mexican people, but I don’t want the US to have Mexico’s economy.
1 Like
1 Like
I do not wish to be Mexico either, but I have a bit more faith in the American entrepreneurial spirit and the resiliency of the American economy. That does not mean we will avoid instability; we will not. But I believe we will weather it, pun intended. 
2 Likes