Being nearer 2050 than 2000 somehow makes the mountain we have to climb to net zero in 2050 harder than it already did. Humanity still seems to be like the frog in the heating pan of water.
Because of so much change happening across political alignment, economic underpinnings, and climate disruption, 2050 is not a prized destination at speed.
> Salesforce, Microsoft, Google all report AI doing 20–50% of work.
It seems likely that this will be true in the coming years. But as a statement about the world today, it would be astounding (if true). What reports are you referencing?
Regarding Microsoft and Google, I've seen statements that AI is writing 20-40% of new code, but "writing X% of code" is very different than "doing X% of work". Other engineering functions (code review, working with other teams such as product management or customer support, etc.) are much less automated, and under the conditions which likely obtain at these companies (established products with large codebases, etc.), AI-generated code still entails quite a bit of manual effort to decompose tasks, generate prompts, review output (which is separate from code review, the latter being done by a second engineer), fix up the AI's output, etc. Not to mention all of the job functions at these companies other than engineering.
Regarding Salesforce, I've seen Benioff's statement that "AI is doing 30 to 50 percent of the work at Salesforce now" and I flat out don't believe it. Has Salesforce laid off 1/3 of its staff? Is it doing 30 to 50 percent more work? Is it somehow a massive outlier from other large companies; if so, why, and if not, then why is virtually no other large organization reporting similar results? It's an extraordinary claim, and requires more evidence than a single comment from the CEO of a company which is in the business of selling these tools.
I believe we are closer to the year 2000, because both positive developments—such as in technology—and negative ones—such as climate change—do not progress in a linear fashion, but rather exponentially, just as the title of your wonderful blog suggests. Therefore, I think—and fear—that the trajectory of the past 25 years will prove more limited than that of the next 25.
There is, however, one exception to this trend of development: our mind. And I am referring not so much to the mindset and culture of the elites, but to that of the average person. Frankly, as someone approaching 70, I see—at least in the Western world—more of a cultural regression than progress. And I don’t wish to enter a political debate about the evolution of more or less democratic, social, or humane forms of government, because pessimism overwhelms me.
Great discussion/overview of your predictions. But I agree with you that the problem was that -you of all people- were too timid and lacking tech confidence to have more challenging predictions. Be brave to be wrong occasionally. Ie. Growth of flying transport dones including for ambulance and other emergency services or emergence of AI level4 (Sam Altman's classification) and some of its impacts.
Uber has 4x the weekly rides of Waymo, and anecdotal evidence is that Waymo is 2x the price. I don't see this happening in this year or even next (though the experience is definitely superior)
Being nearer 2050 than 2000 somehow makes the mountain we have to climb to net zero in 2050 harder than it already did. Humanity still seems to be like the frog in the heating pan of water.
Because of so much change happening across political alignment, economic underpinnings, and climate disruption, 2050 is not a prized destination at speed.
> Salesforce, Microsoft, Google all report AI doing 20–50% of work.
It seems likely that this will be true in the coming years. But as a statement about the world today, it would be astounding (if true). What reports are you referencing?
Regarding Microsoft and Google, I've seen statements that AI is writing 20-40% of new code, but "writing X% of code" is very different than "doing X% of work". Other engineering functions (code review, working with other teams such as product management or customer support, etc.) are much less automated, and under the conditions which likely obtain at these companies (established products with large codebases, etc.), AI-generated code still entails quite a bit of manual effort to decompose tasks, generate prompts, review output (which is separate from code review, the latter being done by a second engineer), fix up the AI's output, etc. Not to mention all of the job functions at these companies other than engineering.
Regarding Salesforce, I've seen Benioff's statement that "AI is doing 30 to 50 percent of the work at Salesforce now" and I flat out don't believe it. Has Salesforce laid off 1/3 of its staff? Is it doing 30 to 50 percent more work? Is it somehow a massive outlier from other large companies; if so, why, and if not, then why is virtually no other large organization reporting similar results? It's an extraordinary claim, and requires more evidence than a single comment from the CEO of a company which is in the business of selling these tools.
I believe we are closer to the year 2000, because both positive developments—such as in technology—and negative ones—such as climate change—do not progress in a linear fashion, but rather exponentially, just as the title of your wonderful blog suggests. Therefore, I think—and fear—that the trajectory of the past 25 years will prove more limited than that of the next 25.
There is, however, one exception to this trend of development: our mind. And I am referring not so much to the mindset and culture of the elites, but to that of the average person. Frankly, as someone approaching 70, I see—at least in the Western world—more of a cultural regression than progress. And I don’t wish to enter a political debate about the evolution of more or less democratic, social, or humane forms of government, because pessimism overwhelms me.
Great discussion/overview of your predictions. But I agree with you that the problem was that -you of all people- were too timid and lacking tech confidence to have more challenging predictions. Be brave to be wrong occasionally. Ie. Growth of flying transport dones including for ambulance and other emergency services or emergence of AI level4 (Sam Altman's classification) and some of its impacts.
I'm skeptical of "Waymo surpasses Uber"
Uber has 4x the weekly rides of Waymo, and anecdotal evidence is that Waymo is 2x the price. I don't see this happening in this year or even next (though the experience is definitely superior)
Blackberry sold far more phones than Apple did once upon a time.
Mark your calendar - say for 30 December this year and come back to this post!
Thank you so much for your clear presentations to we tech dinosaurs!