Yet another Hacker News

15 points surume 2 days ago 29 comments

If you can code practically anything with Claude Code (or equivalents), what types of web businesses will continue to stay viable or profitable?

Source: AI is Killing Saas - https://nmn.gl/blog/ai-killing-b2b-saas

frnkng 2 days ago | parent

Please code me a Microsoft excel with Claude Code.

moomoo11 2 days ago | parent

Any business that solves a problem. The tech is always a commodity. The valuable parts are usually the data, the services offered, etc. and they have good distribution which makes their moat strong.

rvz 2 days ago | parent

Any business that already has a network effect of services and the data that will thrive on even more AI content.

Meta (Instagram, Threads), Bloomberg, X, YouTube, Snap, Netflix, TikTok, Valve.

Coding agents are not designed to clone network effects nor can they.

0xecro1 2 days ago | parent

Been thinking about this a lot myself.

My current answer: go vertical and messy.

Ex, Healthcare portals with ugly data. Compliance platforms with painful regulations. B2B tools with 6-month sales cycles.

lyfeninja 2 days ago | parent

Niche business tools. Ones that solve specific problems or improve workflows unique to industries. I think this is a going to be interesting because you will have a ton of industry SMEs who now can just ask an AI to build them a tool to solve a unique problem they've seen throughout their career and start a business on it (hopefully). I know someone doing this exact thing, and it seems to be working so far.

softwaredoug 2 days ago | parent

Nobody knows

Some hypotheses

(A) AI helps most setup simple tools - even non technical people. But once you get beyond greenfield the cognitive debt builds so you can’t reason about it it. It’s possible mature products don’t get the same gains with AI (or have different types of productivity gains).

(B) We pay a SaaS company to be responsible for an SLA. I personally don’t want to be responsible for that SLA on my vibe coded app so I outsource it. See also support, etc

(C) We pay SaaS to be a reliable source of truth (like Shopify for my Ecom business). The app holds the state of something important. That investment in the ecosystem is itself a moat.

(D) Many “SaaS” businesses are not pure software. They handle payment, benefits, payroll. Often with complex human b2b backends. It looks like just software to us, but we pay to turn a complex set of human relationships into a slick dashboard.

If my business is a pure software tool, it doesn’t have a good moat - and frankly probably nevet did.

milanspeaks 1 day ago | parent

More than 90% for web business will continue to make money.

We are a team of 4 people company and we use 15+ SaaS Web tools and we can code few of them but we see no reason to solve. Why would we replace Calendly already at just $12 per month? Or why would we will create an internal Outlook?

dudewhocodes 1 day ago | parent

I see it the same way. Why would we rebuild a custom solution when $10 per user has it all?

SaaS was always about the service part and works better when other companies use the same tool.

Unique software specific to business needs will get more interesting though.

cyanydeez 1 day ago | parent

Usually its cause the market movers force anyone not moving towards the billionaire VC businness has to raise their cost.

This is usually masked by vertical integrate and buyout of competition.

So either your servicd raises prces to avoid buyout or it gets bought.

RamblingCTO 1 day ago | parent

This is the lifecycle of SaaS tho. It opens up opportunities for competitors to launch and build something more nieche, which is good for customers. And then it repeats

cyanydeez 20 hours ago | parent

Yeah, it repeats until fascism shows up and decides the billionaire class and nepo babies is all "we" need.

BloondAndDoom 1 day ago | parent

Do you think this applies to only cheap software?

I assume expensive SaaS is generally very complicated nothing you can vibecode in a weekend, but I assume if someone is selling $50 SaaS, someone else will vibe code it and sell for $20

Or someone will sell calendly kind of app $10 for a year.

Or maybe everyone will rely on popular brands and this won’t make any meaningful difference in the market other than few categories.

fragmede 1 day ago | parent

Why pay a subscription price even? Maybe we can go back to paying once for a version of a product, and that's it.

RamblingCTO 1 day ago | parent

AI is not killing SaaS, there is no real proof for that. No one wants to host, maintain, and be liable for shit other people can do better for less money. You share R&D costs with other customers. It's more efficient.

This take is so unimaginably stupid and far from the truth it makes me angry. You have network effects, liability, maintenance, mental/managerial complexity, integrations and on and on the list goes. I would be weary of anyone proposing an issue and selling you the solution right with it.

You can also cook your own food. But do you? Do you break your own bread? Do you make your own pasta? No you don't. You pay for an outcome. SaaS is the same. Just because you can do it yourself doesn't mean you should. The code is the smallest part of a SaaS. This will be the same thing like "let's move devs to a cheaper country". Look how that turned out.

chistev 1 day ago | parent

You sound angry, why?

whattheheckheck 1 day ago | parent

No sound was emitted from his typed comment

RamblingCTO 1 day ago | parent

Because these stupid takes are everywhere. It's engagement bait and causes uncertainty with a lot of folks. My LinkedIn is full of that. Just shows me how many people aren't able to think critically and for themselves, blows my mind. And I'm connected to a lot of other CTOs and CxOs ... I think we can do better!

gnz11 3 hours ago | parent

> No one wants to host, maintain, and be liable for shit other people can do better for less money.

I think the point is that AI will be able to do this better and cheaper than the SaaS companies.

al_borland 1 day ago | parent

Thousands of decisions go into building a large SaaS product. Beyond that, it requires infrastructure and maintenance to run it. This can and will break. Even with AI, all of this does not become trivial.

Most companies want to focus on their actual business, which is why they go to SaaS in the first place. They won’t want to do all this stuff themselves, as it’s a distraction from the core business, and they won’t do it as well as people dedicated to building a SaaS product.

raw_anon_1111 1 day ago | parent

All of the major SaaS companies. No company wants to maintain a vibe coded in house SaaS product anymore than they wanted to maintain a Microsoft Access app. AI is a “sustaining innovation” for existing companies.

rithdmc 1 day ago | parent

HackerNews has been saying, for years, that an idea isn't worth anything. Even before AI, there were ten people who could for every person who did release a SaaS side project or business.

The value is in being able to sell it, being able to offset responsibility, all that ancillary stuff, which AI can't do at the moment.

austin-cheney 1 day ago | parent

Not many. The future is self-hosted applications and organization owned/retained data.

raw_anon_1111 1 day ago | parent

Absolutely no company to the first approximation wants to host their own apps and manage updates for a couple dosen different SaaS apps (the industry average).

austin-cheney 22 hours ago | parent

At every step in the cyclical evolution of software as a business people have said the exact same thing: nobody is going to stop doing X for Y. This is always proven wrong even as people continue to advocate for X over Y as Y fully replaces X. Businesses will evaluate this in consideration of expenses and liabilities. Once the Y becomes cheaper than X then X is only retained for emotional reasons.

raw_anon_1111 21 hours ago | parent

We’ve already been through this cycle before SaaS. Why would every company want to go back to the days where you had to have an IT department to manage your apps?

Real businesses aren’t going to over optimize on something that cost $10 per seat.

If it’s “always proven wrong “, is this going to be “The year of Linux on the Desktop”? Is “the iPad going to fail because it doesn’t run Flash?”? Are we going to start browsing the internet on our TVs using set top boxes to get on the “Information Super highway”?

Oh yeah and let’s not forget that the iPod had “less space than the Nomad and no wireless” and was “lame”.

Or don’t forget that Dropbox was going to fail because “For a Linux user, you can already build such a system yourself quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem. From Windows or Mac, this FTP account could be accessed through built-in software.”

I can go on about Dvorak’s comment in 1984 “The Macintosh uses an experimental pointing device called a 'mouse'. There is no evidence that people want to use these things”

Or Michael Dell’s infamous “Apple should just shut down the company and give th money back to shareholders”

austin-cheney 20 hours ago | parent

Remember that data/content is king as the paradigm of business on the web? Its where we have been since the late 90s. Its the advertising economy. Both business and individual consumers don't want to pay for it any more. Self-hosted applications with a cloud premium upgrade are the new business model. For example why bother with all the invasive nonsense of Facebook if a person can self-host something that works better and share it with their real life contacts at their discretion?

The IPod won because it had the Apple app store. People don't need the app store to acquire high quality content any more and they certainly don't want their content locked into the limitations of leased access to some distribution model.

raw_anon_1111 18 hours ago | parent

Right, businesses don’t want to pay $10/seat for SaaS but they want to create server rooms in their offices and hire a bunch of IT people to support it?

> Self-hosted applications with a cloud premium upgrade are the new business model

That’s not happening for any company. Where pray tell are these companies “self hosting” and supporting their business that they are getting a better deal than their per seat licensing cost? Who is supporting their systems?

> For example why bother with all the invasive nonsense of Facebook if a person can self-host something that works better and share it with their real life contacts at their discretion?

So everyone is going to be a server admin and host software and set up their NAT correctly so they can access it via their home internet when their on their phones using their 30 Mpbd upload speeds from their cable modem?

Are people going to start running their own servers now?

> The IPod won because it had the Apple app store

That’s also not true. The iPod was released in 2001, the iTunes Music Store wasn’t released until 2003. As late as 2007 Steve Jobs said that only 5% of music on iPods came from iTunes in his famous “Thoughts on Music”

austin-cheney 3 hours ago | parent

You don’t look for decentralized self sufficiency by looking at random large companies. You find it by looking at the self hosting products and seeing who their paying customers are.

codegeek 20 hours ago | parent

Most B2B SAAS is not going to go to zero just because of AI. B2B Customers don't want to build their own tools generally unless it is their core business. Before AI, customers used to hire developers/vendors to solve a problem. After AI, customers will continue to do the same. They may ask about doing things with AI, sure.

Source: I sell a complex B2B SAAS and even though customers are open to learning more about using AI and are curious, I never had a customer ask so far "Hey so now this AI thing is out, do I really need to you the $xx,xxxx/Year because I can build this in a weekend now, right?". The software is just 1 thing. Customers want stability, support, maintenance, someone to call if shit breaks and if a SAAS vendor does a decent job at it, they won't leave.

If anything, customers who are AI adaptors are asking to collaborate with their existing vendors if possible.