As With Everything, Be Somewhere in the Middle

It may seem obvious, but whenever I’m trying to identify the direction something will go, I try to associate it to similar societal hype cycles. It’s difficult to predict the speed at which ideas will propagate, and mutate, even during a time where everyone is connected. It’s why I see people in small towns fully embracing styles, idealogies, and even personalities that may have peaked in popularity in a major city a decade earlier.

I see such a stark separation around AI sentiment nationally and globally. I went to a comedy show not too long ago and when the comic asked the crowd who uses AI, everyone booed, and the sole woman brave enough to admit it with a hand raise was all but astrocized for it. I realized then I’m in a bubble. YouTube channels denouncing AI are gaining popularity, including Mo Bitar, whom I used to agree with. And although he makes many good points, I fear he’s typecasted himself a touch.

He argues, as many do, that LLMs lack the ability to ever become AGI. And he may be right, but the truth is that LLMs are just one part of a system that is capable of truly revolutionary change. And that’s evident already with Claude Code/Open Code: tool calls, web research, MCPs, and most importantly, frontier models with high agentic scores that can track a list of tasks and bring them full circle after hundreds of steps.

Of course many still don’t know how to use these tools. There are many forms of AI slop in every possible industry, medium, format, etc. We won’t see the end of that soon. As always, approach, opinion, and domain knowledge of that which you use AI on will be the difference between quality and energy waste, because ultimately someone has to use your product for it to be successful.

Circling back to trying to predict where things are going, I now see adamant anti-AI creators as very dated. I think we’ll look back on these years as the infancy that rightfully created doubt, but ultimately cost many the dismissal of an immensely powerful tool.

Even the last six months have monumentally changed my personal experience around using AI. I lead teams across an org that delivers for fortune 100 companies. We don’t vibe code. Ultimately the distinction of quality ends up being the standards that define your product. And as processes mature, tools improve, and most importantly, individuals gain experience and understanding of how AI works, I gain more confidence that we’re at a point where software developers have never had more power to build what they want.

I don’t have AI psychosis. I just have a well documented set of requirements and standards to use as a baseline to test against. And within the last 3 months our output has met those standards at an accelerated rate for the first time. Granted, we have not abandoned practices like reusing frameworks we’ve built and automation. We’re just able to implement these non-AI efforts faster.

I’d love to hear what you think, and whether your opinion has changed.

Also I should mention that my opinion does not excuse the federal and local governments’ behavior around data centers. We need regulation and oversight desperately. I also truly believe the money being poured in still outpaces the value being delivered.. And it’s astounding to me that a MacBook Pro with 64GB of ram can be $5000. Everything that goes up must come down.