I was part of the Internet Revolution.
Back then, the world was discovering what “online” really meant. I led @Microsoft into the Internet Database revolution. Later, I built an Internet company and took it public in 1999.
It was a time of impossible optimism. Every chart pointed up. Every idea seemed inevitable.
We measured value in eyeballs. If people visited, value followed. Revenue was secondary. Reality, optional. When @AOL bought @Time Warner, it felt like proof that the new world had conquered the old one.
But proof can be an illusion. And illusions don’t last.
By early 2000, I had funding from three major sources. All I needed was one. Redundancy felt like safety.
Then April came. The government moved to split Microsoft — the company that had defined the era. That single event was a pin. And when it touched the surface, the entire bubble collapsed.
Pain followed. And pain teaches better than profit ever could.
Now, I’m hearing the same music again. This time, the chorus is AI.
We’re told it will end all jobs. Create universal high income. Birth digital gods by 2027.
I’ve heard this melody before. Twice. In 2000 and again in 2020.
Each time, when truth became uncomfortable, voices rose to drown it out. Not with better arguments — just louder ones.
Three weeks ago, @karpathy Andrej Karpathy said something simple: “Reinforcement learning is terrible.” He added that he’s “semi-pessimistic.” and said predictions will be 10 years not 1.
He wasn’t rejecting AI. He was reminding us that evolution takes time. But the reaction was the same old pattern — utopian shouting to silence the inconvenient. The shouting the past three weeks on AI has been deafening.
For those projecting trillions in value, doubt is dangerous. Confidence is currency. And like all currencies, it can inflate… until it can’t.
I’ve seen this before. Twice.
There is an AI bubble. It will burst. And what comes after will be greater — truer — and far more enduring than what we imagine today.
From experience, David
Enterprise AI Adoption Crisis - 4 Big Issues - A long form essay (Substack)
If you don't agree and feel like enterprise adoption is exploding here is some data and what needs to be fixed. More data and links in the article.
Happy to see some counter views and data though!
@TerryMoran Looking for the right respectful words: idiotic ? Naive? His perspective is looking back and assuming Trump wants the past. For good or bad, Trump is distrusting the post WWII world order. That’s not looking back. It’s a new and dangerous path forward.
Venezuela. Greenland. The Board of Peace. Trump’s @realDonaldTrump actions may be signaling a broader shift in global power dynamics. No opinion here, just observation. Change is in the air. Curious how others interpret it.
@saxena_puru AI software stocks are under pressure because they can never make back the massive investment and none of the problems with Ai, such as consistency, even have a proposed solution. Enterprise solutions are almost non-existent.
@Prathkum Keep preaching it. I agree. Many other reasons, but people like @DarioAmodei are selling the same predictions he made last year and people want to believe. It’s just not true!
So glad to see so many of you waking up to the vibe coding nonsense. @DarioAmodei is selling you. AI cannot replace coders although it makes them massively more efficient. One more time: AI is incapable of replacing a good SWE.
@Hesamation The narrative that ai is replacing programmers is wrong. It’s making them better and more is expected. This will not change and least for many years.
@pmarca is talking about an 80 year tech cycle for AI.
Jensen Huang is saying massive job loss from AI and super intelligence is not even a thing.
Some is finally saying what I’ve been preaching all year.
AI is not ready. It will take years/ decades. Doom is not here.
@mattvanswol I get your point and agree, kinda. The point of sharing should be focused on what will help, “edify” the other person. If it’s just to show off or put them down, then not good. And yes, truth can hurt, but if your intention is to help, that’s good.