Trying to measure the true impact
In attempting to understand the impact of ChatGPT on our world, I have been consistently surprised by the failure of traditional, linear measurements to accurately capture the scope of the changes and effects that appear to be on their way.
Again and again over the last month, I have found that the problem presents itself in orders of magnitude differences, essentially requiring a switch to more of a logarithmic scale to truly grasp the magnitude of the shift caused by ChatGPT's release on November 30th.
The impact of this technology on our approach to informational tasks large and small, as well as the entirely new class of associated "knowledge costs" is significant and unprecedented.
The effects of ChatGPT on our world are just truly staggering, with both possible efficiency gains of 5-10x, or in some specific cases substantially more, as well as the new class of costs required to acquire the knowledge necessary to realizing those efficiency gains mentioned.
Googling is free, but Google's knowledge is quickly becoming a second class citizen to AI.
Linear Scales Not Up to the Task
What I'm finding is that many of the changes simply cannot be accurately measured using traditional linear measurements, and instead require a logarithmic scale to fully understand their impact.
For example, consider a task that used to take 10 hours to complete manually. With ChatGPT, that same task may now be completed in just 1 hour, representing a 10x increase in efficiency.
Some specific tasks may turn multiple weeks of work into just a few hours, a 50-100x increase in efficiency. These increases are simply not captured by traditional linear measurements and require a logarithmic scale to truly grasp their significance.
To be clear, this will absolutely not happen for all tasks. In fact, it probably won't have a direct impact on the majority of the work that folks complete each week. But there are many, many jobs that are likely to be affected, and those affects will have cascading impacts on pretty much everything else, as far as I can tell.
A Whole New Category of Costs
The same is true for costs associated with tasks.
Prior to ChatGPT, many costs were almost non-existent, with costs of $0.00001 being common for a single API request.
"Googling" for information was/is free, but that information is suddenly (imho) in many, many contexts now, an entirely second class citizen when it comes to knowledge.
You an't can't Google follow up questions.
However, with the implementation of ChatGPT, these costs have suddenly become significant, with costs of that have to be measured in pennies, rather than pennies for 1000x the same number of requests. That's 1000% increase in Again, these increases are orders of magnitude greater than traditional linear measurements can accurately capture.
It is clear that the implementation of ChatGPT on November 30th marked a significant shift in the way we approach information work and their associated costs.
While traditional linear measurements may still be sufficient in many cases, they simply cannot capture the full scope of the changes brought about by ChatGPT.
In order to truly understand the impact of this technology on our world, at least in the short term, we're going to need to be comfortable shifting our perspective to adopt a logarithmic scale of measurement in some cases.
Conclusion
The impact of ChatGPT on our world is truly unprecedented and requires a shift in perspective in order to fully grasp its significance. Traditional linear measurements are insufficient in capturing the magnitude of the changes brought, and something closer to a logarithmic scale of measurement is going to be necessary to truly understand its impact on our approach to out daily activities and their associated costs.
The future implications of this technology are vast and it will be interesting to see how it continues to shape our world.
No comments:
Post a Comment