Hacker News Reader

Comment by cardosof | Hacker News Reader

It's been around 8 years since this happened: I started learning R and a Gartner consultant told me I was wasting my time because, "as Gartner puts it", the future of data science is citizen data scientists with very little coding skills because the tools and frameworks will do all the heavy lifting. I ignored him because I liked the language anyway.

I guess every now and then someone will try to sell a tool to "build nerd things without the nerds!" be it web dev, data science/engineering, ux, devops, whatever.

I will believe in it only when I see it: everyday fellas building sustainable saas businesses on their own, without coding.

I'm less skeptical towards little internal applications though. Those don't need to be good as standalone products, they just have to be better than shitty forms/sheets and cruds companies use internally. In that space I think Retool is doing something cool and interesting.

View on HN

If someone can build a low/no-code product, then the barrier to entry is so astonishingly low and there will be a ton of competition for that product. Another consideration is that these low/no-code products almost always cost money.

So not only will the market be flooded with clones, rent-seeking from the provider will eat into margins.

Much like the cloud is someone else's computer; SaaS is someone else's programmer.

It's not like the barrier to learning basic Python is that high. Or Excel macros.

The “someone else’s programmer” metaphor isn’t totally accurate, because of a key difference between cloud vendors and no/low-code vendors.

To support more customers, a cloud vendor needs to have linearly more computers. A no/low-code vendor, on the other hand, does not need to have linearly more programmers.

Generally I share your skepticism, but a lot of BI tools are low-/no-code. Generally you can't build a SaaS platform on them any more than you can build it on Excel.

That's a great point, thanks.

I deal with Microsoft’s PowerBI and it is true that somewhat normal people can do a lot but the generated result ends up extremely inefficient and we usually end up getting sql experts in to manually write the queries anyway because it costs an obscene amount for the resources to run these auto generated reports.

Looker is the best example IMO.

Don't bother with "no code" but make it a lot easier for people to write the code/markup and also much easier to edit en masse (one sed command in a Github repo, rather than editing every single dashboard separately in Tableau/PowerBI).

This. I've just built out an internal CRM for my travel company using Airtable+StackerHQ.com and for my business, it's hands down better than anything commercially available... because it covers all the nuance of our individual situation.

2 months ago, I was a bit down on no-code, but the experience on building on Stacker was good enough to change my mind. Softr.io is also getting there, too.

> because it covers all the nuance of our individual situation.

Without wanting to downplay your achievement, I've worked for organisations that did the same thing with MS Access back in the early oughts.

The 'custom' application grew with the organisation. Features were bolted on over the years inelegantly. It was impossible to integrate easily with other systems.

The quirks of the system were known by long time employees who developed work arounds that were convoluted and took time.

UX wasn't a thing back then; but if it had been people would have definitely said it had terrible UX.

We've been promised these solutions for a long time. In my experience, they overpromise and underdeliver.

overpromise and underdeliver.

Something that under-delivers is much more useful that something that doesn't exist. In most situations and most companies the the choice is never between the ugly Access thing and top quality software written by team of highly educated professional programmers. It's a choice the between the ugly Access thing and an ad-hoc collection of text files and post it notes.

> Something that under-delivers is much more useful that something that doesn't exist.

No. I disagree.

Sometime a badly constructed solution is worse than no solution.

> It's a choice the between the ugly Access thing and an ad-hoc collection of text files and post it notes.

I think this was definitely true before the advent of SaaS products.

I understand your point. The reality is that no solution will ever be perfect, be it an off the shelf SaaS or my homebrew no-code build.

There are a few things here which help us;

- Airtable has all the integrations a human being could ever need. Our stack is Airtable + Stripe + Zapier + Wordpress + ActiveCampaign.

- General awareness of the concept of UX. You need a product owner who fiercely defends the product. And .. you need a company culture which treats employee satisfaction as something that actually matters. I know that these things not common.

- I have no viable alternative. SaaS companies in the tour & activity space struggle to make money and struggle to invest in their product.

- A community of contractors. Airtable has a large community of able freelancers and contractors who can come help us out, if needed.

Thanks for your comment, I just looked at AirTable and StackerHQ. A quick question: is there a way to export your data to standard file formats, just in case you wanted to move off the platform to other vendors or write everything yourself?

I think that getting started quickly on a system that you will likely never have to leave makes a lot of sense, but it is good to have an exit strategy. Like with Heroku: keep running small projects there, but being able to move to alternatives.

Yes, Airtable is exportable. That's part of the reason I chose this solution. I'm big on redundancy.

As a wise investor once told me, “once it shows up in Gartner, it’s done innovating.”

I think low code tools have come a long way. I also think the problem with low code tools is the specific problematic edge cases or implementation cases that make it hard to do without the more robust capabilities of actual code. That, and the baseline that if you can produce a low code program, you can code, for which the barrier of difficulty is constantly lowering anyway with more and more robust libraries and tools.

I didn’t catch anything in the piece about development that could be a revenue-producing product. Rather, sounded like internal apps that might have been done in Excel or Access plus some VBA goodness back in the day.

There was a bubble.io app (I tried bubble, didn’t enjoy it) built by a non techie, that is making 30k per month in revenue (I am trying to remember the name). Some of these no code / low code tools are really getting powerful and easier to use, at least for CRUD apps that aren’t complex.

I am guessing there are needs for tons of simple tools and businesses that programmers aren’t aware of, or not interested in working on. These would be met by savvy subject matter experts whose tech skills are good enough to build something with these tools.

Ever tried Azure Customvision? Building image classification / object detection models is trivial using this platform. I was highly surprised by the accuracy and low cost... I mean anyone who can upload and tag images can make a production ready image classification model at pretty low cost...

I see Excel and PowerApps everywhere across our customers.

Reply on news.ycombinator.com