I do believe that if we are heading down this path of almost total automation, the infrastructure should be public and not owned by corporations. Perhaps corporations could build on top of it to bring innovation and such. But leaving it in the hands of single individuals is too much power in one hand, and we have far more history and lessons to show us what happens when power is concentrated and not shared.
I feel that passivity is upon us. We are push to use a tool that most people don't really know how it works, others perceive has a sentient being. When this is attention economics working as designed.
Even recent news from Microsoft new AI Agent that addiction is one of the goals establish by senior management.
It's the systems and the people behind these systems we need to look for.
I feel sad when I read content about people being supposedly replaced by LLMs, the provocative and sometimes not true on how someone fired an entire department just to use a £200 subscription app.
The low regard for people some individuals have for is concerning.
If we were indeed in the presence of an artificial intelligence machine, why can't it understand the world? Why does it need tremendous amounts of data just to distinguish a bat from a bee?
This is a deep learning construct we should be using it as a means to explore possibilities, not as truth.
I think the people who say LLM AI can replace humans everywhere are, possibly deliberately, missing the point. Used properly AI enhances human talent and makes humans more efficient. That may replace some humans, but it also allows other humans to do more with less and therefore allows them to create useful product and make a living they otherwise couldn't. So in terms of human employment I expect AI to be a wash eventually. That doesn't mean it won't be tough for the people who see their job/career path AIed away, but I don't see that as much different to the general computer use removing swathes of clerical workers in the 1980s/90s . In fact what we saw out of that is that a real Personal Assistant (as opposed to a note taker with that title) is invaluable for a senior executive and that people who managed to remain PAs had a far more interesting and responsible career than they might otherwise have had. I expect much the same to apply to AI.
Of course all this also depends on people using AI correctly and I strongly suspect that the people talking about firing workers to replace them with AI are using it wrong. That's going to come back and bite them. I discuss some of this here - https://ombreolivier.substack.com/p/ai-bicycle-or-train?r=7yrqz
I agree with you that AI works best when it supports people and helps them do better work. That is where I see the real value is.
Where I see it differently is in how companies behave. Many organisations focus on short-term cost savings, not long-term value. Because of that, they may try to replace people too quickly and use AI in the wrong way.
So while some companies will use AI well, others will damage quality and only realise it later.
The risk is not just losing jobs, but being in companies that make poor decisions.
I feel good outcomes depend on good leadership, and that is not always present.
Yes. And in fact we have already seen companies regret some mass firings and attempt to rehire the staff they fired. Needless to say this has not been easy and obviously doesn't lead to a positive relationship between the top level managers and the workers.
It doesn't help at all. It does show to us where some companies values are. And how little they have us in regard. Automation has help us in so many ways, but I feel some must not be implemented, just because we can doesn't mean we should.
Frankly, I think the AI business model needs adjustment through Policy. Humans need their privacy protected where as the current AI model relies on scooping-up as much human info as possible to re-mix it and rent it back to humans as 'digital smarts.' I'm not certain how accurate language translation is with AI, but that is also part of the challenge with AI output - fact checking.
First, most modern AI models aren’t literally scooping up personal data in real time they’re trained on large datasets, often a mix of licensed, public, and synthetic data, then it uses other tools to gather data across the web. That said, the industry has not been transparent enough about what goes into those datasets or how consent is handled, and that’s a real policy gap.
Second, on outputs, these models doesn’t know facts it predicts likely sequences based on patterns. Translation is actually one of the stronger use cases, but even there, nuance and context can break down, so human verification still matters.
Trust is the core issue. Regulation will help, but better engineering practices, clear data boundaries, auditable systems, and accountability for how AI is applied are just as important.
I feel this is something we cannot let be in the hands of companies.
I get what you say, although we need to be hopeful, and understand that in these systems, certain voices are amplified. And true it feels we get use to define ourselves to our job titles, and now that some companies and individuals feel we no longer are necessary, are trying to taking away what most hold dear.
We're in the early stages of powerful but limited tools. Let's use them to augment human potential, not replace human value.
What's your experience with AI in the workplace? Are you seeing augmentation or replacement?
I do believe that if we are heading down this path of almost total automation, the infrastructure should be public and not owned by corporations. Perhaps corporations could build on top of it to bring innovation and such. But leaving it in the hands of single individuals is too much power in one hand, and we have far more history and lessons to show us what happens when power is concentrated and not shared.
I think we need a better operating model.
That means clearer goals, smaller claims, honest measurement, and leadership that focuses on outcomes instead of spectacle.
If AI is used well, it should make people more effective.
If it is used badly, it becomes theatre.
Most of the noise today is not about capability. It is about narrative, power, and short-term pressure.
I feel that passivity is upon us. We are push to use a tool that most people don't really know how it works, others perceive has a sentient being. When this is attention economics working as designed.
Even recent news from Microsoft new AI Agent that addiction is one of the goals establish by senior management.
It's the systems and the people behind these systems we need to look for.
I feel sad when I read content about people being supposedly replaced by LLMs, the provocative and sometimes not true on how someone fired an entire department just to use a £200 subscription app.
The low regard for people some individuals have for is concerning.
If we were indeed in the presence of an artificial intelligence machine, why can't it understand the world? Why does it need tremendous amounts of data just to distinguish a bat from a bee?
This is a deep learning construct we should be using it as a means to explore possibilities, not as truth.
I think the people who say LLM AI can replace humans everywhere are, possibly deliberately, missing the point. Used properly AI enhances human talent and makes humans more efficient. That may replace some humans, but it also allows other humans to do more with less and therefore allows them to create useful product and make a living they otherwise couldn't. So in terms of human employment I expect AI to be a wash eventually. That doesn't mean it won't be tough for the people who see their job/career path AIed away, but I don't see that as much different to the general computer use removing swathes of clerical workers in the 1980s/90s . In fact what we saw out of that is that a real Personal Assistant (as opposed to a note taker with that title) is invaluable for a senior executive and that people who managed to remain PAs had a far more interesting and responsible career than they might otherwise have had. I expect much the same to apply to AI.
Of course all this also depends on people using AI correctly and I strongly suspect that the people talking about firing workers to replace them with AI are using it wrong. That's going to come back and bite them. I discuss some of this here - https://ombreolivier.substack.com/p/ai-bicycle-or-train?r=7yrqz
I agree with you that AI works best when it supports people and helps them do better work. That is where I see the real value is.
Where I see it differently is in how companies behave. Many organisations focus on short-term cost savings, not long-term value. Because of that, they may try to replace people too quickly and use AI in the wrong way.
So while some companies will use AI well, others will damage quality and only realise it later.
The risk is not just losing jobs, but being in companies that make poor decisions.
I feel good outcomes depend on good leadership, and that is not always present.
Yes. And in fact we have already seen companies regret some mass firings and attempt to rehire the staff they fired. Needless to say this has not been easy and obviously doesn't lead to a positive relationship between the top level managers and the workers.
It doesn't help at all. It does show to us where some companies values are. And how little they have us in regard. Automation has help us in so many ways, but I feel some must not be implemented, just because we can doesn't mean we should.
Frankly, I think the AI business model needs adjustment through Policy. Humans need their privacy protected where as the current AI model relies on scooping-up as much human info as possible to re-mix it and rent it back to humans as 'digital smarts.' I'm not certain how accurate language translation is with AI, but that is also part of the challenge with AI output - fact checking.
You’re right to question this.
First, most modern AI models aren’t literally scooping up personal data in real time they’re trained on large datasets, often a mix of licensed, public, and synthetic data, then it uses other tools to gather data across the web. That said, the industry has not been transparent enough about what goes into those datasets or how consent is handled, and that’s a real policy gap.
Second, on outputs, these models doesn’t know facts it predicts likely sequences based on patterns. Translation is actually one of the stronger use cases, but even there, nuance and context can break down, so human verification still matters.
Trust is the core issue. Regulation will help, but better engineering practices, clear data boundaries, auditable systems, and accountability for how AI is applied are just as important.
I feel this is something we cannot let be in the hands of companies.
We must protect our cognitive sovereignty!
Yes we must, and is not that difficult has it seems. Practice ethical reflection, civic awareness, and digital literacy not just technical skills.
It's scary when we read about, we need to abandon ourselves to be more like a cold machine. And worst how recently being human feels like a burden.
I get what you say, although we need to be hopeful, and understand that in these systems, certain voices are amplified. And true it feels we get use to define ourselves to our job titles, and now that some companies and individuals feel we no longer are necessary, are trying to taking away what most hold dear.