There’s little question among the tech savvy that AI will have a profound impact in coming years. Yet, AI’s supposed impact is often lost, even on some smart people in tech.
To that point: One engineer with decades in the aerospace industry, including patents he’s been awarded, told me he heard a futurist speak about AI at the IPC APEX 2025 event in Anaheim last week and felt blindsided by the personal revelation he experienced—despite knowing about generative AI for years going back to Open AI and ChatGPT in late 2022.
This is roughly what he said: “This keynoter was going on about how half of the APEX audience—we are mostly engineers—was approaching AI like we approached the PC in the 1980s. Half of us didn’t use PCs for years, and I thought: ‘Yeah, that was me.’ He went on to play a song he created with AI about the event just by giving it clues. And gave other examples. It was amazing.”
Of course. Turns out the keynote was by Kevin Surace, a talented and energetic keynote speaker who has delivered a similar message to other audiences. A brief interview with Surace is online.
Most tech journalists would not care about his keynote message at APEX because Nvidia GTC 2025 was happening at the same time in San Jose. Over at GTC, CEO Jensen Huang pointed out at least a dozen hardware technologies to be enlivened with AI software including future GPUs like Blackwell Ultra and Rubin.
RELATED: Nvidia announces Blackwell Ultra, Rubin roadmap plans at GTC
Huang is very sensitive to making Nvidia’s message clearcut, but he and so many CEOs in tech have fallen victim to rapidly changing innovations. He was also speaking at GTC to his supporters and customers and insiders who have been in the know for years, so he gets a pass on speaking fast and with many slides and a glossary of abbreviations and acronyms. He’s also speaking to investors who know his competitors as well as he does.
We have the AI intelligentsia and we have everybody else.
Still, there’s an interesting divide happening in the AI realm, shown by that keynote at APEX happening at about the same time as Jensen’s keynote. We have the AI intelligentsia and we have everybody else. And “everybody else” probably includes millions of people with great IQs or even engineering degrees, including materials and mechanical engineers who never needed to code to do groundbreaking work. Everybody else also includes the policymakers in Washington, mostly, who took forever to work up a CHIPS and Science Act. Smart people don’t like to show their weak side, obviously.
Probably, Surace made an impact because he understands his audience, with a broad range of backgrounds and abilities. It reminds me of having taught many classes of freshman composition at two community colleges and a big urban university. In such situations, you might expect to have students arrive at an entry college level at least having written a short essay, but immediately I found out that 25% had no concept of a thesis sentence. A small minority could not say what components comprise a full sentence. Another small minority could start the semester already knowing how to write an effective 500-word essay. So, a pretty wide range of skills.
This provides an insight to what’s happening with AI. Some large organizations are beyond testing AI and now deploying it, broadly, but have not revealed much about the benefits. Others are using an outside service provider to help. Meanwhile, a vast array of workers are fiddling with tools from Open AI and others, without much coordination by management. I have interviewed coders using various tools to write a rough draft of code, checking for accuracy and enhancing it with added prompts. Entire companies are cropping up to coach us in writing prompts for making AI insights more productive. Meanwhile, even the terms we use are confusing to people like the engineer above, who called the prompts being used “clues.” (A much simpler and, arguably, clearer term.)
At Fierce, we have used AI to draw images, prompted by a human.( Fierce Electronics does not allow an article to be published written solely by AI, but it is a good starting resource for basic research. ) College deans I’ve interviewed are using AI to write references and summaries, mostly as a time saver. It makes sense to try out various tools. Lately, I’ve used Grok, accessed in X.com, not as a political statement, but mainly because it is so simple to use. It draws politically-incorrect political cartoons, for example, and takes some wacky prompts and offers surprising results that revealed insights I had not have thought of on my own.
The point of trying out AI in a simple way (like asking it to create a short poem in iambic pentameter on an upcoming event) is to gain confidence and insight for what is coming with AI in the next decade. This effort might reduce anxiety about future jobs being lost, but it could inspire greater insight and deeper thought. As Elon Musk and Bill Gates and others have said, we are facing a time when AI will soon become the main tool we rely upon for work and creativity. It will also mean we have to create a social structure that accommodates entire societies where work is redefined and how daily resources like food and housing are accumulated by needy people even if there is not a work product created that’s attributable to a GDP. (Ugh, I was trying to avoid calling it a form of socialism or ‘on the dole’ but we still don’t have a good vocabulary for the ancillary impacts of an AI society. And, what would Musk as head of DOGE think of admitting there’s socialism on the horizon as he cuts thousands of government jobs?) Fortunately, all that is years off—but not too many years off.
How are we to get ready for the really big questions of 2040 and later if we can’t prepare even the smart folks for what AI means? It means a lot of work and also recognizing the work birthed by AI is actually kind of immense and far beyond whether Nvidia can create a megawatt server. I recall as an instructor how I could indeed fail many students, but my real job was to get them to practice and re-write, re-write until they were a C level or above in less than 12 weeks. That’s similar to the job the AI industry faces on a mega scale.
What do you think? Write me at [email protected]. And it might be fun to see whether you believe the entire freshman composition enterprise could be avoided entirely if we just accept that AI can handle writing 500-word essays with a few prompts.