10 ms·

Even with their current capabilities, these AI systems will dramatically improve productivity. We could have a 20-year AI 'winter' with no new advancements, an

Even with their current capabilities, these AI systems will dramatically improve productivity. We could have a 20-year AI 'winter' with no new advancements, and they'd still be a big deal. The thing is that it will take years to integrate them into existing domains and workflows. Honestly I think the most relevant comparison is the rise of desktop computers themselves. Suddenly paper-based processes had to become electronic and in some cases that took 40 years.

lm28469 2y ago
> Even with their current capabilities, these AI systems will dramatically improve productivity It's been 24 months, where is it? And even what you can measure doesn't match the valuation of openai at all
- lifty 2y ago
  You won’t necessarily notice it as much in end user application. But I already see good adoption for knowledge workers and the next step is in process automation through agentic workflows. Most likely the result will be in the form of efficiency gains, improved profits and worse customer service.
  - apwell23 2y ago
    why hasn't all the productivity boost reported in qtrly results from corporates.
tbrownaw 2y ago
> Suddenly paper-based processes had to become electronic and in some cases that took 40 years. "Took", past tense?
threeseed 2y ago
Except we've had AI systems for a while now and it hasn't meaningfully impacted productivity across the economy. Maybe in select pockets e.g. knowledge workers and even then it's highly debatable. You compare this to the internet or smartphones and they have been transformative across every aspect of society. And as someone who works for a bank which is heavily exploring LLMs there is no complexity in integrating them into existing workflows. The issue is that (a) risk of privacy/security being compromised through prompt exploits and (b) risk of reputational damage if the prompt is biased or hallucinates. Issues that may well be inherent to all transformer based models.
- jncfhnb 2y ago
  Since you’re saying “for a while” I presume you mean ML. ML systems have had an enormous impact on productivity. There’s a huge volume of decisions that get made instantly and with far greater precision using models That were done manually before.
  - pydry 2y ago
    I've worked with ML for a while and I wouldnt say that this is the case at all. ML also didnt replace manual decision making. A lot of previously automated decision making was done with what were basically encoded rules of thumb which didnt overfit much worse than "more advanced" ML models did.
    - jncfhnb 2y ago
      Your personal failure to use ML successfully has no bearing on its wider deployment and success. Overfitting strongly implies that you or whoever was doing it just didn’t know what they were doing. ML is used everywhere.
      - pydry 2y ago
        Wow, salty :) How many ML projects for large businesses have you been on?
        jncfhnb 2y ago
        Dozens. I’m a consultant and it’s what I do. The modeling is very easy. The operational changes are the hard part. Basically any random task of. A sort of maintenance, production, replacement, procurement or scheduling task can save 10-30% with a simple model. There’s a long way to go but this stuff is everywhere. The world hasn’t even fully realized the productivity value of spreadsheets and email.
        m_rpn 2y ago

lm28469 2y ago

> Even with their current capabilities, these AI systems will dramatically improve productivity It's been 24 months, where is it? And even what you can measure doesn't match the valuation of openai at all

lifty 2y ago
You won’t necessarily notice it as much in end user application. But I already see good adoption for knowledge workers and the next step is in process automation through agentic workflows. Most likely the result will be in the form of efficiency gains, improved profits and worse customer service.
- apwell23 2y ago
  why hasn't all the productivity boost reported in qtrly results from corporates.

tbrownaw 2y ago

> Suddenly paper-based processes had to become electronic and in some cases that took 40 years. "Took", past tense?

threeseed 2y ago

Except we've had AI systems for a while now and it hasn't meaningfully impacted productivity across the economy. Maybe in select pockets e.g. knowledge workers and even then it's highly debatable. You compare this to the internet or smartphones and they have been transformative across every aspect of society. And as someone who works for a bank which is heavily exploring LLMs there is no complexity in integrating them into existing workflows. The issue is that (a) risk of privacy/security being compromised through prompt exploits and (b) risk of reputational damage if the prompt is biased or hallucinates. Issues that may well be inherent to all transformer based models.

jncfhnb 2y ago
Since you’re saying “for a while” I presume you mean ML. ML systems have had an enormous impact on productivity. There’s a huge volume of decisions that get made instantly and with far greater precision using models That were done manually before.
- pydry 2y ago
  I've worked with ML for a while and I wouldnt say that this is the case at all. ML also didnt replace manual decision making. A lot of previously automated decision making was done with what were basically encoded rules of thumb which didnt overfit much worse than "more advanced" ML models did.
  - jncfhnb 2y ago
    Your personal failure to use ML successfully has no bearing on its wider deployment and success. Overfitting strongly implies that you or whoever was doing it just didn’t know what they were doing. ML is used everywhere.
    - pydry 2y ago
      Wow, salty :) How many ML projects for large businesses have you been on?
      - jncfhnb 2y ago
        Dozens. I’m a consultant and it’s what I do. The modeling is very easy. The operational changes are the hard part. Basically any random task of. A sort of maintenance, production, replacement, procurement or scheduling task can save 10-30% with a simple model. There’s a long way to go but this stuff is everywhere. The world hasn’t even fully realized the productivity value of spreadsheets and email.
        m_rpn 2y ago