Rui CarmoI am most definitely not the first person to mention this, but the past year has seen a Cambrian explosion of two things: AI deniers who base their judgement on very limited exposure (or effort to use it) and thousands of variations on software of all kinds, from the perennial to-do list to AI-infused toothpicks.
Which, let’s face it, is a bit much. It’s bad enough to go on social media and be pitched things I don’t need, but the hypefest around new startups hawking their wares to fill hitherto undiscovered “problems” in increasingly pointless sub-niches is a bit much too. I find the entrepreneurial energy heartening, but most of them solve no real problems whatsoever.
I mean, how many code editors can you have anyway? And workflow engines? And CRM pipelines? There is a dizzying amount of almost-identical AI-driven solutions out there trying to land the ideal five-word VC catchphrase, and it’s just tiresome at this point.
Plus if you take the long view, most of them are completely divorced from the realities of, say, traditional IT, which has turned into an ouroboros of complexity thanks to geological eras of non-technical decision-makers – but that’s beside the point right now.
The one thing that I find positive among all this (and that I’ve hinted at a few times in the past) is that anyone with a systematic take on things can now create whatever tools they might need to scratch their particular itch(es) – the results may not be perfect, and they may have an audience of one, but they can unblock time, patience and capabilities they would otherwise lack and improve things in general for themselves, without having to conform to someone’s idea of a “product” or jumping through hoops to make their workflows fit a tool.
Of course, the flip side is that it is all too easy to go overboard (when you have AI, everything looks like a solvable problem), so restraint and taste – which is probably the most effective, even if somewhat passive-aggressive and snobbish criticism you can lob at people these days – are essential.
But this ability for people to design tailored solutions for themselves is, I think, what SaaS vendors (and system integrators) fear the most. In an age where anyone can point an LLM at a bunch of APIs and say “build me a CRM that works for me” (which is something I’ve seen done, to the horror of a friend of mine working at Salesforce), the “value” of customisations has to be replaced by something, and believe me, it’s not going to be a consulting firm’s business acumen – or even their taste (which is nearly always lacking).
That does not make the tools magical, of course – they still need proper boundaries, mature tooling and someone who understands the result.
But on a smaller, personal scale, it’s kind of ironic to think that probably the main reason this is going to kill the appeal of most SaaS is that we’ve spent a couple of decades killing native desktop development in favour of kludgy, half-baked web UIs that are all variations on lists of things people click on.
And believe me, if there is one thing that LLMs can do perfectly, it’s web interfaces with lists of things people click on.