Don't Call it a Bubble

Lately, there’s been a lot of talk in the tech world about how we’re in an AI bubble, with many drawing parallels to the dot-com bubble of the late '90s. The argument goes like this: AI is hyped up beyond belief, and investors are throwing insane amounts of money at startups simply because they’ve slapped the “AI” label on their product, regardless of whether it has a compelling vision or a solid product behind it. And you know what? Those concerns aren’t entirely unfounded.

There is a ton of hype around AI right now. We’re seeing an influx of AI startups, some of which don’t seem to have much going for them beyond a fancy pitch deck and a vague promise of AI-driven magic. And yes, the money pouring into these ventures is staggering. But here’s the thing: Just because there’s hype and some questionable investments doesn’t mean AI, as a whole, is “a bubble.”

Real Innovations in AI Are Already Here

Let’s get real for a second—AI isn’t just about buzzwords and sky-high valuations. Over the last couple of years, we’ve seen some groundbreaking advancements in Generative AI (GenAI) and Large Language Models (LLMs) that are genuinely transforming how people work and what they can accomplish on their own.

I’ve been fortunate enough to work on some incredible products like Luster that really push the boundaries of how AI can help make people more successful, more efficient, and more capable of tapping into their ideas. This product lets sales teams practice realistic conversations without the stress of losing a real customer. We spent a lot of time polishing the experience to ensure that each persona brought unique challenges to the user and that every conversation could be different — offering what will likely be a record for replayability in the SaaS space.

The creazy thing is, I wouldn’t even consider that a complex AI use case, and it’s doing more than most people think is possible with AI right now.

AI is also opening new doors for all types of people around the world. For example, tools like Opus.pro are a game-changer for content creators, allowing them to take long-form videos and break them down into shareable shorts with just a few clicks. This isn’t just saving time; it’s making high-quality content creation accessible to people who don’t have a full production team behind them. Pair that with something like Runwayml, and services like Elevenlabs, and anyone with a great idea can create engaging content in video, audio, interactive, and written formats with the help of AI. It removed barries in technology, education, and skill in a meaningful way.

The Bubble Claim: A Lack of Depth?

Here’s where I think the “AI is a bubble” argument falls flat. A lot of the people making this claim haven’t really built anything substantial with AI. Take the tool I’m writing in right now, beehiiv, for example. They’ve integrated AI writing and image generation into their editor, but in my experience, these tools are practically unusable.

I’ve tried generating content with their built-in AI, and I’ve deleted or undone almost everything it’s produced. Including this weird and creepy image of a bee I tried to generate to make my point for this post. You can compare to the image generated under it with the same basic prompt thrown into GPT 4o.

If your experience with AI is limited to the generic chat experience that most people see through LLMs like Claude or ChatGPT, then yeah, I can see why you might think it’s all hype. These are powerful tools, but they still operate with a garbage in > garbage out “rule”. This means the base models we’re working with need some context and direction to generate something useful like humans do.

Not everyone needs to be an AI expert, but if you’re going to claim that AI is “a bubble,” you should at least understand what AI is capable of and how the overall technology is being used.

AI Is Disruptive — But It’s Not All About LLMs

Let’s be clear—there is a lot of noise in the AI space right now. Not every AI startup is going to be the next big thing, and some will inevitably fail. One of the things that motivated me to write this post, was reading one from the beehiiv co-founder, Tyler Denk where he said,

Which was prefaced by

Tyler was actually addressing some irresponsible spending and trends in the startup space that I totally agree with, but this wasn’t the first time I’ve heard someone claim that it’s a bubble.

Dismissing AI as a whole because of a few bad actors or overhyped products is shortsighted. Will some people lose a LOT money by investing in something without substance? Absolutely. Will some people make insane amounts of money by investing in companies that will literally change the way the world behaves, works, and thinks? Yep. I don’t know if those things will even out, but the “eventually” in the post I quoted about is really where my caution to you is rooted.

So who do I trust or listen to?

I’ve been spoiled by working with some incredibly smart engineers at EIG and the AI team at Automattic before that. If you don’t have a direct line to those types of people, here’s what to look for:

• Foundational Knowledge: They understand the history and underlying technologies, like machine learning, natural language processing (NLP), and the hardware requirements (e.g., GPUs) needed to power AI systems. This filters out people who just started learning about AI in the last 1-2 years and who think that GPT IS AI.

• Experience Levels: While they may have valid input, casual users experimenting with tools like ChatGPT aren’t experts. Look for individuals who are deeply engaged with AI, like data scientists, AI engineers, and product developers who actively build and test multiple tools. Even if they aren’t building products, you should look for people who are actively tinkering to find ways to leverage AI for themselves.

• Active Involvement: Credible AI practitioners are hands-on, regularly experimenting with new AI tools, understanding the intricacies of prompting, and working with specialized models and use cases beyond text generation. Check out the HuggingFace community to get a sense of who the really active leaders are in this space.

• Diverse Applications: They have a broad understanding of AI’s impact across various industries and applications and can help explain what types of AI or AI implementations make sense for different use cases. I’d rather talk to someone who challenges how I’m thinking about it, than someone who just agrees with every idea I have.
  

• For example, my most recent team was great about challenging whether or not AI was the right solution based on the problem the client was trying to solve. Sometimes ML is a better fit, and sometimes good old fashioned software engineering does the trick.

• Continuous Learning: They keep up with the rapid pace of AI advancements, constantly updating their skills and knowledge to stay ahead of the curve. There are huge announcements, new research, and interesting releases coming out every day so you should be getting advice or input from people who know where AI is at right now, not where it was at 3 months ago.

My Game Changing AI Tools and How I Use Them

To wrap things up, here’s a quick look at the AI tools that have either totally changed how I work, made meaningful impact on my productivity, or enabled me to do/product more than I could before they existed.

• Perplexity.ai: My go-to for initial research, understanding customer segments, and answering complex questions—whether it’s work-related or my kids’ latest curiosity about space travel, it saves me hours of time searching, vetting, scanning, and processing information daily. Plus it’s very shareable and easy for anyone to understand.

• Opus.pro: Perfect for transforming long-form videos into shareable shorts. The manual process of editing a video down to clips is the reason I didn’t get into creating videos earlier in my career. I knew the opportunity was there but the time investment never made sense.

• GPT Plus: I use GPT-4o for quick tasks, but I’ve also created a suite of custom GPTs tailored for specific needs, like generating images for my newsletter or analyzing audience data from Sparktoro.

• RunwayML: My go-to for creating eye-catching background videos for social media, saving time and resources on video production. They recently released a new version of the text-to-video model and people (and brands) are creating some incredible content with this platform.

• Canva and Krea.ai: Canva handles most of my design needs, while Krea.ai is for those projects that need a bit more “wow” factor. I plan to dig deeper on Krea, but every release they put out feels like magic to me — especially having been an early user of Photoshop.

• Reclaim.ai: A lifesaver for managing my schedule, automatically blocking out time for deep work, decompression, and those little tasks that often slip through the cracks.

What tools are you using and how? What’s your take on AI? Reply to the email and I might include your tips, ideas, or opinions in a future post.