I gave a talk last week called [Democratizing Research Platforms](https://medium.com/@caseyg/democratizing-research-platforms-7b8f177b50db) foreshadowing a near future where even the deepest enterprise b2b SaaS might be replaced by internal platforms, open-source, and/or some form of on-demand generated tools. I received [a great set of questions in response](https://www.linkedin.com/feed/update/urn:li:activity:7239657064040443904?commentUrn=urn%3Ali%3Acomment%3A%28activity%3A7239657064040443904%2C7239739570370535424%29&dashCommentUrn=urn%3Ali%3Afsd_comment%3A%287239739570370535424%2Curn%3Ali%3Aactivity%3A7239657064040443904%29) from a librarian friend: Are you really saying I — a librarian — can write my own software? And I’m _already busy_, do I really have time for this nonsense? An interesting headline from a few weeks ago: Klarna, the payments provider, is shutting down their Salesforce and Workday instances in favor of [generating internal tools](https://seekingalpha.com/news/4144652-klarna-shuts-down-salesforce-as-service-provider-workday-to-meet-same-fate-amid-ai-initiatives). Questionable leadership decisions aside, these are the big kahunas of SaaS being shown the door. So it does feel to me like an inflection point in the hype cycle around SaaS. B2B enterprise SaaS is like a buffet where you can get a little bit of [any food that somebody could possibly want], but none of it is particularly good, and you pay by the pound. Oftentimes a business even pre-pays for empty plates that they estimate may be filled, as if they were renting the shitty buffet out entirely. And to be fair, a buffet is an expensive and complicated operation to run, and to keep stocked and staffed. However, if we look at a real use case for “food”, we can narrow the scope of complex operations down really quickly into something much more manageable. For example: * you are a vegetarian * it’s lunch time We have just immediately ruled out meat and dinner/breakfast. With this (simple, basic) level of context and understanding, we can actually get you a _better_ lunch **for your needs** than what you’d get at a buffet. And it will be a lot cheaper than the buffet too, where the target customer is [the lowest common denominator of all hungry people in this town at any given time of day]. [todo: Pictures of home cooked meals next to their culinary counterparts] You don’t need to know how to build a stovetop to fry an egg. But that’s the logic through which software development is gatekept as the exclusive right of computer nerds — whether self taught, bootcamped, or degreed. There is some risk here. Cars, for example, have become so computational that it’s no longer the domain of any odd mechanic to help with repairs. Many cars now require authorized service by the manufacturer, which is more “fixing bugs” than repairing. (The movement against this broader trend is called “right to repair”, meaning that if you break it, you should also have the right (with the proper tools, skills, and determination) to fix it. In today’s world of consumer products this is often not the case.) There are different schools of thought on whether people understand and can meet their own needs. An apocryphal quote: If we asked the people what they wanted, they would’ve said a faster horse. And it’s true that if you ask a language model for a faster horse, it’s going to give you a faster horse — not a car. So in some respects we are entering into an incredibly literal, compliant, and even uncreative or anti-innovative new mode of human-machine collaboration. But I find the idea that users are incapable of understanding their own needs and building systems to self-actualize their goals to be a quite cynical technocratic worldview. So the other question: I’m already busy — even given a hypothetical toolmaking machine, do I really have time? Is it worth it? A lot of “innovation” is not all it’s cracked up to be, and is essentially cover for repackaging the same crap with some added bells that customers didn’t actually want, need, or ask for. (This brand of innovation is increasingly widely known as “enshittification”. And manufacturing desire for these un/non-critical and useless enshittovations(?) is the fulcrum of consumerist societal growth and/or the business model for b2b enterprise SaaS.) Another example: If I’m hosting a Thanksgiving dinner, I don’t actually need (or want) an Internet Of Things-enabled smart oven™️. (Though I unfortunately own one. Sad but true!) What I could use and do want is three more burners. Bunsen burners, even! This kind of low-art zero-confabulation but tailored to specific local or even individual needs “innovation” is where what Maggie Appleton describes as “home-cooked software” might find its sweet spot. Product talks a good game but most productivity workflows are not and do not need to be rocket science or high art (as construed by “design engineering”).