On not keeping up with the AI hype

Engineer Dave argues that developers should not feel pressured to keep up with every new AI tool or model, as the hype is partly driven by vendors' incentives. He advises focusing on delivering value with current tools and maintaining a healthy macro-level awareness through team collaboration or curated social media, rather than burning out on the bleeding edge.

Like many of you even the ones who won’t admit it , I have a bad habit of scrolling Instagram for a bit when I wake up. I “fixed” it by doom scrolling X and LinkedIn instead. Genius move - apparently reading about strangers shipping side projects at 5am lol, make that 7:45 makes you feel measurably worse than watching Insta-slop-reels. Over the past 2-3 years roughly 27 AI-years we’ve gotten used to having to keep up with a new model every month or so. That’s fine, manageable even. Next you must keep up with the newest IDE or TUI, there’s a new one every week. Then you have the new agent harnesses and “must have” skill sets every 12 hours - and you are loop-maxxing aren’t you ? If not, pfft… might as well brush up on your welding skills and find a new career </s None of these individually is a problem. Tech has always moved fast just ask any Javascript developer . The novelty and extreme-FOMO seem to have combined into this anxiety-inducing AI-fever many are experiencing. Cynical-Dave thought: the whiplash is partly structural - everyone shipping these tools has an incentive to make “keeping up” feel mandatory. Keep that in mind. If your job is building reliable systems, and you’re doing that today, then keep doing it. AI is a tool. You need macro-level awareness of where the industry is going - the same way you’ve always needed it for languages, ORMs, clouds. You didn’t personally evaluate every JS framework in 2016, yet the ones that mattered still found you when it mattered. If you’re an engineer who is effectively using the tools available to you to deliver good products to your users/stakeholders/customers, and you’re on average enjoying your job, then you’re sweet as https://www.facebook.com/watch/?v=847668813999072 . Similarly, as an engineering manager or VP of Engineering, are your teams happy? Delivering value against their OKRs/KPIs/goals? Nice work Keep slaying queen, as my kids say. But using the latest model which is 12% better btw trustmebro to summarise your meetings likely won’t move the needle If you work with people who each individually have a healthy interest in staying up to date without burning themselves out then, between you, you’ll have a great macro-view of the industry. Your collectively distilled sense will likely be just as effective as one person spending 60 hours a week ingesting every article and testing every tool. You’ll be happier, and likely better at your day-job. If you work solo, or in a very small team, spend some time curating your social media to cut out the noise as much as you can. Engage with user groups in your area, chat to people in your co-working space. “Have you tried this new Sonnet model?”, “How are you validating what your agent s build?”, “I’ve just tried switching from Codex to OpenCode. What do you use?”, etc. Don’t worry about sounding stupid. This is a very common reason I hear for people not discussing this stuff more openly; “everyone else seems way more advanced than me, and I don’t want to embarrass myself”. Here’s a super-secret pro-tip that 9/10 dentists agree on; nobody knows what the fuck they’re talking about. There is no single best way to do your job, and anyone who says there is also has some amazing snake oil to sell you. I still doom-scroll through the various platforms, and like this morning still catch myself freaking out that I’m not doing enough. So this is just as much for me as anyone else. I think the point is that very few people in the world actually need to keep up with the bleeding edge - regardless of the tech we’re talking about. Always keen to discuss if you have thoughts on this, especially if you disagree and see other patterns or requirements. Cheers, Dave