Median product "lifespan" has fallen from ~4y down to ~1.5y

Anecdotally, I've found I'm far less interested in adopting new products and rethinking my workflows amidst a sea of half-baked and short-lived products / services.

Instead, I'm stripping back to basics, building small tools to solve specific problems, and find myself vastly preferring established software with a track record of keeping feature-bloat down and the problem-space focused.

Me too. Even personal task planning. Dumped Trello for lists in plain text mails. Accessible and editable everywhere (I just forward the changes to myself).

I decided that I don't want a private life that needs GANTT charts or org mode for interesting projects. After half a dozen of time management books I realized that this is just another form of procrastinating: * reading a lot about time management * overcomplex tooling * shuffling time slots in a calendar

The solution is a simple (check) list as a reminder. Calendar only for hard deadlines. Plus "just do it" attitude. The rest will follow. Or not, if life is otherwise too busy -- which I simply accepted.

The results don't differ much from before. Except: less time wasted with software.

I run a stack with garbage collection for most of my tasks: I just keep things I need to do in my head, and if I think of it a lot it has references to it. If I stop thinking about it it will eventually be purged by GC :D

LLMs have flipped the build vs buy debate on its head for me. I also have been re-evaluating existing tools and building my own custom, just-for-me versions and it's been liberating. Using off-the-shelf software increasingly feels stifling and limiting. I do wonder if the future for many SaaS companies is to provide a solid core with a nice API/MCP/Hooks/etc to build on top of so you can add features only your company cares about instead of every SaaS trying to be a one-size-fits-all.

The cool part is building smaller tools to solve your problems also allows you link those all together (MCP/CLI/API) in ways that weren't possible in the past using LLMs as a glue layer.

With regard to ‘provide a API/etc.”, companies have been trying hard to do that for a long time, even coming up with their own languages, like salesforce did.

It’s extremely hard to do, even on a basic level. And the worst part is actually the UI. You can go the Wordpress/shopify route where you allow client templates and it’s a huge cost for you to maintain. Or you can do some sort of automatic field generation and no-one’s happy.

Not only that, but your product ends up in a straight jacket, where you can’t fix any fundamental data structure mistakes because half your client integrations will break.

> Distribution and marketing have become much more important than before - within the context that everyone can build "features" for free.

Nothing about this statistic supports this claim. Maybe, actually, features aren't free, and even with a glamorous HN marketing blitz, nobody wants to use the dogshit software LLMs produce. Everyone can build "features" for free, if "features" is an alias for "a pile of 100 bugs with some makeup on it to make it look like a feature if you squint at it from a distance".

Death of the sunk-cost fallacy, you launch a thing and nobody gives a shit especially you yourself.

However I think there will be a price to pay building software without the "trough of sorrow" phase stressing over the product and market fit and limping away with lessons learned - the way you build software and the problems you solve are basically confined to a safe space, a bubble of yourself + your AI agents and a brief moment of suspense submitting to reddit or whatever, and you risk learning nothing at all to improve your next projects and higher-volume won't help that.

I certainly agree that product lifespans are getting shorter, but despite the very common belief that everyone can build products overnight, I still think building a good product is very, very hard.

The reason is that most people try to skip the fundamentals and create a demo overnight, assuming people will want it. We had more than 200 conversations with customers to truly understand what they were looking for, and we had to evolve our offering significantly along the way.

I don't think building a good product overnight is realistic. AI gives you the ability to build much faster, but it's far from free, considering the cost of tokens. I agree that it's much cheaper, though. What it doesn't give you is the knowledge that comes from understanding your customers and going deeper into a problem they actually want to solve.

I also agree with you that distribution and marketing will be key. One thing to remember is that you should start with marketing and distribution, not with the product. If you spend all your time building the product first, it can be very hard to sell it later, and you may not have enough time to find your audience, especially if you're bootstrapped.

I'd recommend focusing on marketing first, building second, and getting real-time feedback from your target audience throughout the process. Talk to your customers every single day and keep learning from them. That's something AI cannot do for you on its own.

The idea that users "make their own tool with AI" is a nerd fantasy that does not exist in the real world. Some use cases get absorbed into products like ChatGPT because Jim no longer needs a special web-based tool/calculator/product and can just ask ChatGPT to do it for him. But normies can't and don't want to build or maintain products. They wouldn't even have the patience to write the prompt to ask the AI to build it for them. Vibe-coders are mainly wannabe programmers and nerd hobbyists.

Software is in its SoundCloud rap era.

This could signal a healthy fail fast competitive market. Maybe?

Fail fast, with no one succeeding is not healthy.

One vibe-coded app is no better than other, or if it's better - not for long. So why would anyone invest in product that is going to be superseded in a few months/weeks/days?

This is the median lifespan. If the number of total products is increasing, that doesn’t suggest that fewer products are succeeding. You could even have more products succeeding, if the number of failures is growing even faster

The problem I'm seeing is not with vibe coding itself, although it is absolutely the LLMs that empower people to fall into the actual trap: Failing to do any research whatsoever into their market and subject area. They see one surface-level problem and instead of digging into the problem space to gain a deep understanding of what would make a solution good or bad... they just code. Vibe code, code by hand, somewhere in the middle, whatever - the current trend is to code first and ask questions later. Some people always did that, but these days it seems like everyone who has not built products before is doing it.

We're seeing it more and more. Someone whips out an app independently, launches it, then says, "How do I find users?", "Why does nobody use my app?" With the answer always being the same - they skipped the step of actually talking to potential users before coding.

So we see what is being described - lots of throwaway projects.

This creates a negative feedback loop. Developers release a product and if they don’t instantly get millions of users, the project is abandoned. This teaches users not to bother with new software and services that won’t stick around, making it even less likely that the project will survive the first year.

I used to sign up for new stuff all the time, but I don’t bother anymore. I don’t trust anything will still exist in 6 months.

With TV shows I have a soft rule that I don’t bother starting a show until it’s at least 3 seasons. I’m thinking of taking a similar approach in software, where I won’t bother with anything that hasn’t been actively supported and updated for at least 3 years. If the developer isn’t committed to the idea, why should I commit as a user or customer?

While there are certainly benefits to lowering the barrier of entry to starting a company or releasing software. I think the lack of commitment and seriousness on the part of those releasing it is a major drawback for users. In a brick and mortar business, owners put a lot on the line to get up and running, they expect not to make a profit for several years, and fight like hell to make it work. Meanwhile in modern software, it’s just a bunch of people throwing spaghetti at the wall, hoping something with stick long enough for a 7+ figure exit.

Software used to solve problems for users. Now AI is solving most problems for users, but not all (although ever increasing). I think software needs to evolve to solve the category of problems that AI is not (yet) good at. For example very hard (extreme long horizon) problems in digital space which are still expensive to solve. Some domains overlapping with real-world and restricted/regulated industries are also not yet disrupted by AI to the level regular SaaS is. To get ahead one needs to start in those domains now, as it will get much harder once AI will start eating them.

I find myself pushed to develop tools of my own because even market incumbents can't meet my niche use cases for certain things.

In the same vein, I am trying not to feel dismayed as I venture into the SaaS landscape in a bid to decouple myself from employment and you cannot help but feel your product is doomed before it even launches given people can build their own solutions.

You just need to remember, not everyone is as technical as the people here and those individuals still require solutions to problems.

It's tiring hearing people say "oh well I'll just build it instead" yeah but is it maintainable? Does it scale? Do you know how your data is handled and saved?

However, my ability to produce a lot of really useful software for myself has gone way up. None of them are over a year yet but I use them more than the non personalized apps they replaced

Not SaaS, but I had an idea for using a CPU emulator to run a compiled binary under Java, with libc implemented in Java so standard file IO works. I punted on the idea because it seemed pretty difficult and maybe take six months. I recently gave Claude a Java RISC V emulator and a partial bridge between Java and libc, and in under an hour, it could run bash and sqlite from Debian.

I have no idea how correct the code is or what I'd do with this, and I have no plans to publish it for those reasons.

I would 100% agree on that. Now REA comes in again and in a few years everybody could clone big and complex software. I think a software will need more than features in the future.

It will need licences, or content others could not deliver. Really something you can’t clone that fast.

I think about Netflix, Spotify, Amazon. You can clone all this platforms but will never have the rights, artists or the logistics for example.

I would 100% agree on that. Now REA comes in again and in the future everyone can clone complex software.

I think distribution become more important than ever. And having something you can’t clone on day 1. I think about Amazon, Netflix, Spotify.

You can clone the platforms but not the licences, artists logistic etc…

Could just be composition. nothing launched in 2025 or 2026 can show up as 4 years old yet, and there's way more of them now.

I think another element the cost of maintaining it, supporting it, marketing, etc. hasn't been commoditized like producing brand new software. Those things take time to do properly and there are lots of people churning out a SaaS and then they get abandoned because its not fun anymore to do all the administrative/sales stuff.

I seriously don't think non-technical people are pumping out apps, probably never will. They simply don't have an interest and would rather just pay for a service. A HVAC, plumbing, or electrician company isn't going to write their own CRM system. Now they will price hunt those services so as software production becomes cheaper, I would expect SaaS margins to also be going down.

Wow, that makes Google products seem supported for ever.

Testing products, getting a feel for them, thinking about them, talking and exploring, I mean this is still there and it's long been the actual time-consuming part of making great software; where the soft skills intersect with the hard.

Maybe the technical stuff is more accessible then it was beforehand but the thousands of hours of the other stuff that leads to good software is still thousands of hours.

Quality has become less work, but only by a little.

But this doesn't mean people value the craftsmanship because they never did before...

It's always been simply a popularity contest and now it's that more than ever.

The question is, why are you spending your time building things? Is it to make things that are finely crafted? Is it the road to wealth? As a tool to get other work done? Just make sure you have an answer.

It's your time after all...

Distribution and marketing have always been more important. The hard part has always been being consistent, not having a half-baked prototype.

Frankly AI is making people even lazier than before, which means that anyone willing to stick around and be consistent is probably the winner.

Now that barriers to entry are going down, I think software will become a hit and fad driven industry similar to movies, music and video games.

Certainly not for infrastructure.

Arguably there have been fads for infrastructure, just on a longer time scale.

> a hit and fad driven industry similar to movies, music and video games

That doesn't track at all.

When the movie and music industries were pumping out their biggest hits, the production cost was high. Profit may be higher now in the modern era while costs are lower, but would you honestly (seriously honestly please) call most of the remake/homage slop we've had since the late 2000s "hits"? This simply isn't survivorship bias either. Just look at what music stands the test of time on youtube and spotify or gets people at least smiling in real life. Look at the variety of genres too. Look at when we had the highest density of solid audience ratings for even run-of-the-mill or cult movies.

You can say all you want that these industries "always did this" and insist that when you look at your own list of favorites it's just because you're getting old, but that's more obviously denial as the years of this bullshit have racked up. Even the artists agree even when they're finding success at this game, and they have the same resentment. Younger people don't even care about movies and music anymore while older people desperately pay and hope for new stuff like it's another pull of the increasingly rigged slot machine. It's not even that younger people are more interested in social media. That excuse is just as clearly false when you see how hard they go on retro styles and what they also have saved on playlists.

"hit driven" in the business model sense.

The movie studio business model depends on putting out a whole lot of stuff and having a few big hits earn enough to cover the losses from the flops.

Now that software is basically "dollars in, software out", companies will probably put out a lot more software. Similar to movie studios, most of this software will not hit, but the few things that do hit can earn back the costs of the things that do not.

I don't know. I still think you're looking at this from the consumer side exactly the way they want you to.

In business, some of the most reliable returns come from predicting when the wheels fall off and riding that sucker until the bitter end. The "hits" don't matter anymore to that kind of business. The area under that tail part of the curve can be much larger than the area under the earlier peaks when consumers were most enthusiastic.

it's always been like that. how far back do we want to go? mootools versus jquery? flash? foxpro and vb? cp/m and dos?

software has always been a popularity contest

this website itself is vibecoded slop, a quick look at the listings shows hallucinations.

Interesting, thanks for posting