Hacker News Reader

Comment by ldjkfkdsjnv | Hacker News Reader

People also dont realize that this will kill SAAS margins. If anyone can quickly build working software, the whole moat is gone.

If you dont think this will not happen, you have not used cursor.sh with sonnet 3.5. I am working on a startup that is effectively a copy of existing products. My flow goes:

1. Screen shot competitor UX

2. Drop screen shot into cursor.sh, ask it to code it (there will be issues)

3. Screen shot individual components in the above screenshot

4. Have cursor.sh fix each component one by one

5. Go back through and tell it each issue by highlighting the code

6. Iterate a bit

7. 90% of the UI has been copied in one day, with a working backend.

This is with a next.js/supabase app deployed on vercel. For complex B2B saas.

Its over. In a year the models will get so much better

View on HN

1. Install wordpress 2. Install theme

It’s over! All websites can now just be slightly tweaked templates. Go home, we no longer need any web dev whatsoever.

It’s not like the bar will raise. No, everyone will expect the exact same level of quality as of 2024 and bloated, AI diarrhea using 200x resources is fine now and will always be fine.

Nobody needs to differentiate and we will all copy each other endlessly and eventually the whole software world will be basically one single product.

The collective backlog of every software company probably stretches into centuries.

This gives us the tooling to A.) burn through those faster. B.) solve more complex problems.

If your entire job as an engineer is “take tickets with very clear acceptance criteria and turn into component / basic business logic”, I’d be worried for sure. I was initially worried myself when all this landed two years ago, but using it daily and now building enterprise scale apps with it, I know we’ve got at least another decade or two before it’s able to tackle the full breadth of an engineers role.

This sounds like the software engineering equivalent of setting up an Amazon business account to get into drop shipping. IMO if it becomes easier for any random person to dump AI generated SAAS apps it will only strengthen the current big players.

Remember how Brick and Mortar was dead because of online shopping? Who knew in 2024 I'd be more likely to head out to an actual store than risk purchasing some sketchy product off Amazon, but I can imagine that if we do eventually actually get to a future where the world is filled with the white noise of SAAS garbage coming out left and right anyone actually interested in mitigating risk for their business is going to learn quickly to err on the side of caution.

If anyone and everyone can emit noise then that becomes the baseline, and people will catch on quick to find those that rise above.

>This is with a next.js/supabase app deployed on vercel. For complex B2B saas.

I have seen this said before. Would you show some of these sites that you have knocked out in one day fully working? If not, would you be willing to do the community a service and over the course of a weekend, live-stream or video yourself duplicating this site[0]? Just the publicly available parts.

It would be eye-opening for many I think, and in the process would make you one of the most famous developers in the world (or at least in the US), since at least two very well-known developers/influencers do not think it is currently possible and dunking on them live or on video screen recording would really be a sea-change in our sector. Not to mention, settling once and for all the "LLMs are going to take our jobs" vs. "lol, that are just a fancier autocomplete" argument within the industry.

I am not joking about this or trying to bait you, I would seriously like to know if I am wasting my time with traditional development and really need to re-orient my skills and job prospects. So far, I have seen many people do demos, but none were anything other than relatively simple, error-prone, SPA apps. It is more complex ones, I am interested in. The link below[0] is not that complex for a traditional developer, but it isn't super-simple either. From your comment, it sounds like you have the experience and ability to do it.

[0]https://neetcode.io

One important thing to remember is the cost right now is not being passed to you. I wonder how this all changes when companies are forced to charge you the actual costs vs being heavily subsidized.

While this is true, I would bet that the cost is less than hiring a software engineer in the United States

My experience with this tech is you’re not nearly as able to solve problems because you didn’t actually write the code. So now you need to be better at reading code you didn’t write. The models can’t do a number of common programming tasks, but they are indeed super useful.

This tech currently just makes each engineer more efficient. There’s an infinite space of problems to solve. I guess we’ll must be digging through that space faster and wider from now on.

This is about making a quick buck with throwaway products and deceive customers into thinking this product has any future.

Besides, any sufficiently complex field has a lot of business rules that you can't derive from screenshots.

The level of delusion regarding AI makes me want to completely ignore anything that touches it and come back in 10-20 years when people learned something.

Maybe OP meant "all CRUD TODO-like SaaS is toasted".

I recommend trying it out. Just don’t drink the Kool-aid.

What happens if you have to maintain and extend the codebase over many years? Will your workflow work for fixing complex bugs reported by users?

SaaS type software has been so easy to build for so long and the margins have remained strong. I think the reason for that is that the SaaS software is an artifact of defining exactly what is needed for a particular function and that remains valuable and expensive

Is anything you've made by this process publicly available?

This sounds a lot like programming, from 10k feet.

Interesting! So is the marketing and sales process going to be even more of a differentiator than it is now?

In the short run yes, but i have no idea what happens in the long run

Copying has never been a hard problem.

Yeah but you couldn't copy major functionality in one day. The same can be done from a figma. This will bleed into every area of software

A big part of tools is hiding complexity from you - how does this workflow handle that?

What if you want to make things that are new?

Just put it in Figma. Its not that much different. This is going to kill software margins, any piece of saas can be quickly copied

Poor on-call guys

Reply on news.ycombinator.com