You're Not Behind
Why I stopped chasing every new AI tool
Yesterday it was Clawdbot. The day before, some paper that would “change everything.” Next week, it’ll be something else.
If you’re in tech and paying any attention to AI, you know the feeling. You open LinkedIn and fifteen people in your feed are posting about the same tool. Same screenshots. Same breathless enthusiasm. Same phrases: “silently released,” “no one is talking about this,” “this changes everything.”
The irony, of course, is that you’re only seeing these posts because everyone is already talking about it.
I’m writing this as someone who feels the FOMO too. When Clawdbot started flooding my feed yesterday, my first instinct was to drop everything and try it. I haven’t yet. The idea is genuinely interesting. But when I ask myself what I’d actually use a personal AI assistant for, I don’t have a good answer. And I suspect most of the people posting about it don’t either.
The Feedback Loop
A small group of AI influencers follow each other closely. When one posts about a new tool, others pick it up within hours. They add their own angle, their own take, but it’s the same underlying content rippling outward. By the end of the day, your entire feed looks like coordinated marketing.
These are people optimising for followers, not for helping you make good decisions about where to spend your time. That creates a skewed perception of what actually matters.
Here’s the question I keep asking myself: how do these people have time to discover every new tool, test it thoroughly, and write thoughtful posts about it? They don’t. They’re playing a different game entirely.
The Graveyard of Hype
Remember prompt engineering? For about a year, it was going to be the job of the future. Companies were hiring prompt engineers at six-figure salaries. Anthropic was advertising roles at $375K. LinkedIn profiles with “Prompt Engineer” in the title surged.
Then LLMs got better. The models started understanding what you meant without needing carefully crafted incantations. By 2025, job postings for prompt engineers had essentially disappeared. The skill got absorbed into broader roles. The hype moved on.
This isn’t ancient history. This was eighteen months ago.
Or consider the AI hardware moment of 2024. The Humane AI Pin launched at $699 plus a $24 monthly subscription. It was going to replace your smartphone. The reviews were brutal. Slow, overheating, underwhelming. Returns outpaced sales. By early 2025, the company had sold its assets to HP for a fraction of what it had raised.
The Rabbit R1 followed a similar arc. 100,000 preorders based on slick demos. Then people actually used it and discovered it was, in the words of reviewers, “a solution looking for a problem that doesn’t exist.”
Go back a bit further. NFTs. The metaverse. Web3. Billions of dollars poured into these trends. Ninety percent of Web3 projects failed. NFT trading crashed 93%. Metaverse platforms like Decentraland maintained only 379 daily active users despite billion-dollar valuations.
The hype machine has a short memory. But the patterns repeat.
The Fundamentals Evolve Slowly
I used to focus on prompt engineering. Now I focus on context engineering. The shift wasn’t about chasing a new buzzword. It was about recognising that as models improved, the bottleneck moved.
In my work, I’m constantly dealing with data that doesn’t fit neatly into a context window. SOPs, Excel sheets, call transcripts, process documentation. The skill isn’t knowing which model to use. It’s knowing how to structure and load the right information at the right time. Which sections matter for this specific task. How to transform messy operational data into something a model can actually work with.
These techniques transfer regardless of which model or tool you’re using. GPT-5.2, Claude, whatever comes next. The tools change. The thinking compounds.
Let the Dust Settle
I’ve developed a simple heuristic for dealing with the firehose: when something new gets hyped, I note it and move on. I don’t drop everything to try it. I don’t feel guilty about not being an early adopter. I wait.
If it’s still around when the next hyped thing comes out, then I dig deeper. Most things aren’t.
This feels counterintuitive in an industry that celebrates speed. But 99.9% of people are still figuring out how to use basic AI chat applications effectively. If you’re in tech and using Claude Code regularly, you’re probably in the top 0.01%. You’re already ahead by far.
The anxiety that you’re falling behind is manufactured by a content ecosystem that profits from your attention, not your outcomes.
The Real Work
There’s no doubt AI is redefining the tech industry. The capabilities are genuinely transformative. I see it in my work with clients every day.
But the transformation isn’t happening through whatever tool launched yesterday. It’s happening through the slow, unglamorous work of figuring out how to apply these capabilities to real problems. Understanding your domain. Building trust with stakeholders. Extracting operational logic from people who’ve never had to articulate their tacit knowledge.
None of that requires you to be on the bleeding edge of every new release. Chasing the bleeding edge often distracts from the work that actually matters.
The hype machine will keep churning. New tools will keep launching. Influencers will keep posting about things that will “change everything.”
Most of it won’t matter. And that’s okay.
You’re not behind.