I wouldn't invest in open-source companies, even though I ran one • Hacker News Board

Re: I wouldn't invest in open-source companies, even though I ran one

#161

Post by XCSme » Mon, Jan 25, 2021, 12:57 PM UTC

apple4ever wrote: post #139

XCSme wrote: post #123

Great post, I completely agree! I have debated over this for years while building a self-hosted analytics platform (userTrack.net) and I have always reached the same conclusions: if I open-source it, not only I affect its sustainability but also the business goals would become complete opposite to the product goals (eg. as you mentioned, making it easier to install improves product but affects the business). I now th…

I think the pricing you mentioned is also part of the issue. Companies that are open source but paid something have a crazy high price for that something. I know they are going after the Enterprise market but they still price themselves out.

I also found out the same thing, but it's mostly related to paid software in general. It's a lot easier to sell expensive software. The lower your set the price the more "cheap" customers you will have, which results in more support tickets, maintenance, marketing needed, etc. The customers who want lower prices are also those who would be more critical of your product, as the price they paid for it means more to them compared to someone who doesn't mind spending $50/mo for a basic service.

I am still trying to find a way to accommodate for both enterprise customers and low-budget customers. I tried doing this using some simple licensing terms (eg. get this if you are a solo dev, get this if you are large business), but people will just buy the cheapest option that has the features they want. This makes it hard to offer the same features at different prices and might lead to either downgrading the basic product or adding some buzzword features for the enterprise product.

Re: I wouldn't invest in open-source companies, even though I ran one

#162

Post by thu2111 » Mon, Jan 25, 2021, 1:01 PM UTC

sitkack wrote: post #145

totoglazer wrote: post #141

Earlier quoted context omitted.

This is fascinating! Could you share some names of these databases? I’d love to learn more.

BigQuery and Snowflake are my top two.

Snowflake is based on FoundationDB which was proprietary for years but is now open sourced, developed by Apple. So it's maybe more like open core than totally proprietary.

Re: I wouldn't invest in open-source companies, even though I ran one

#163

Post by thu2111 » Mon, Jan 25, 2021, 1:03 PM UTC

tinco wrote: post #114

mooreds wrote: post #103

Earlier quoted context omitted.

This is a really good point. However, there are still successes that I'd never have expected. Jetbrains is a good example. Who would have thought that an IDE could be sold when there are adequate ones for free? There are also many many smaller niches where software can be sold because there isn't adequate developer interest to create OSS. One example is this: https://querix.com/products/lycia (not a customer, just ra…

I did, there's never been really good community driven open source IDE's. For every community driven IDE I can name you an enterprise driven IDE that's popular and arguably better. Code editors are a product where polish has high value, and users are willing to pay for that polish. I don't particularly care if my editor is open source as long as it's got strong community and great polish.

IntelliJ is open source. It was once proprietary but these days is "open core".

The Ultimate edition is the open source version of IJ with more plugins. However the rebranded non-Java versions like Goland aren't open source, unless you consider the IJ base. They are the core language plugin(s) + integration and rebranding code.

Re: I wouldn't invest in open-source companies, even though I ran one

#164

Post by thu2111 » Mon, Jan 25, 2021, 1:06 PM UTC

tinco wrote: post #117

hapless wrote: post #90

The author makes a good case against open source businesses, but he indulges himself in a false dichotomy between a traditional proprietary model and open source. Increasingly often, there is no market for your proprietary software. It is then not a choice between open source and proprietary models, but rather between doing open source or doing nothing. A lot of markets are just gone , and they're not coming back. Od…

Exactly, and open source companies make great money too. ElasticSearch being a prime example, yes AWS eats a large share of their pie, but who cares, the pie is absolutely huge. Another example is GitHub and Gitlab. Both got big by leveraging the open source community for marketing. Imagine all the resources GitHub spends to enable open source, that's a cost center that will never go away. If they hadn't done it thou…

Is GitLab profitable? They seem to have $120M in 2019 revenue but I'm not sure how big they are.

GitHub got big but then sold to MS. I'm not sure it was ever big in terms of profits. If the argument for investing in open source companies is that they can get a big exit by selling to a proprietary software company then I'm not sure the argument is strong.

Re: I wouldn't invest in open-source companies, even though I ran one

#165

Post by monoideism » Mon, Jan 25, 2021, 2:13 PM UTC

jabl wrote: post #150

monoideism wrote: post #106

Earlier quoted context omitted.

To be clear, I don't typically have any issue with corporate use of my code. That's why I chose MIT and other permissive licenses. That said, what I had in mind when I licensed my code were start-ups and small+medium businesses. It didn't occur to me that behemoths like Amazon might stripmine open source and fail to contribute back. Even though the permissive licenses permit this kind of behavior, it seems exploitati…

> Far more importantly, it also failed to occur to me that tech companies would become the new railroads -- massive monopolies that dominate society, the economy, culture, politics, etc. These are social and political problems, you can't fix them by sitting in your basement writing copyleft code anymore than you fix them by writing permissively licensed code. It's not like the FAANG's of the world would go, oh no we…

> These are social and political problems, you can't fix them by sitting in your basement writing copyleft code anymore than you fix them by writing permissively licensed code.

This is true. But just because you can't fix something by yourself, doesn't mean that you should contribute to, or participate in, the problem.

Re: I wouldn't invest in open-source companies, even though I ran one

#166

Post by mooreds » Mon, Jan 25, 2021, 2:40 PM UTC

lifeisstillgood wrote: post #119

mooreds wrote: post #79

Earlier quoted context omitted.

>FOSS short circuits this gatekeeping function There are other ways past gatekeeping: * free trials * money back guarantees * free as in beer solutions Open source goes one step further than those and says "you don't just get to use this, you get to own it" where "own" means change as you like.

two of those are either need a purchase order upfront, or in 30 days (which is waaay shorter than the time it takes to actually raise a PO!) and ... who gives free as in beer but not open source software? I mean winzip is all that comes to mind? But it does not change the essence - only FOSS gives you software "forever" whilst jumping over the otherwise immutable accounting department t - it is probably the greatest…

I think OSS is a tremendous distribution hack; no argument there.

The issue arises when the developers want to get paid so they can eat/pay their mortgage.

Yes, sure, you've jumped past the PO process and your software is popular. But now you have to crawl back into it so you can get paid.

And the best ways to get paid are laid out pretty well in the original post:

* consulting

* hosting

* open core

and are all problematic for different reasons.

Re: I wouldn't invest in open-source companies, even though I ran one

#167

Post by ignoramous » Mon, Jan 25, 2021, 3:31 PM UTC

jrochkind1 wrote: post #81

ignoramous wrote: post #52

If engineers are the most important resource, I think being a company known for open source might act as an inbound sales for attracting top talent, too (see: Google, RedHat, Mozilla). For engineers, it makes sense to work for companies that do have FOSS projects because they can always point to their corpus of work when switching between jobs as proof of their worth. Most engineers who realise this don't ever want t…

There's possibly "companies with FOSS products" that arne't "FOSS companies". For instance, I know people that work for heroku that work on open source (and are primary maintainers in some case) on company time -- but heroku is obviously not a FOSS company. Or, in a comment below jedberg says 'At Netflix our ethos was "if it's involved in infrastructure, open source it, if it's related to movies, keep it closed.' --…

> This seems to be less and less the case. Because companies are more ruthlessly focused on the bottom line, and don't think they can afford to have anyone working on something they don't own or that might help a competitor? I dunno.

For one, if one's open source product is so good that BigTech wraps its software around it, then that's a cause for validation and celebration. For example, it is pointless for the Rust community to bemoan Google's use of Rust to improve ChromeOS' security; or for Dr. D. Richard Hipp to bemoan pretty much every company for building money making software around SQLite.

Besides, like Bryan Cantrill points out in the blog post I linked to from above, it is often trivial for BigTech to expend effort to re-implement the API surface (MongoDB -> DocumentDB) or even the rewrite whole software (MariaDB -> Aurora) if that's what they need to do to compete with anyone's FOSS / source-available offering.

> There's possibly "companies with FOSS products" that arne't "FOSS companies".

Of course, I am not saying FOSS is the one true way when just recently tech companies like Zoom and Snowflake have demonstrated the potential of proprietary tech; but just that FOSS will commodotize your core, may be half-heartedly, or even half as effectively, but it will happen and there are business opportunities to be built around it. In fact, FOSS is one of the last few avenues left to take on BigTech, as far as technology is concerned.

When I say FOSS, I don't mean the kind of throw-all-commits-over-the-wall, superficial-community-engagement FOSS Google does with Android and Chrome. Google can do so because they have got the money to keep funding "an open source loss leader". For any other upstart it'd be suicidal to keep such an iron-grip on their FOSSd codebase (like Elastic will find out in the coming years). The right approach for such upstarts is to create a Foundation around the tech (k8s) or donate to one (Cloudera / Hortonworks), and generally try to really rake in the benefits that being open-source brings.

Re: I wouldn't invest in open-source companies, even though I ran one

#168

Post by nemothekid » Mon, Jan 25, 2021, 6:19 PM UTC

enriquto wrote: post #147

nemothekid wrote: post #77

Earlier quoted context omitted.

This is a post about running an open source company. In other words, you have to worry about your prospective customer's lawyers objecting to your product just because it is in part AGPL. You could just upsell with a different license; but from a business model perspective that makes the distribution channel of open source irrelevant.

What you are describing sounds more like a problem with sales, or marketing, than with legal. I don't think your legal department would prefer you to use BSD instead of AGPL.

>I don't think your legal department would prefer you to use BSD instead of AGPL.

Google's legal department is the problem: https://opensource.google/docs/using/agpl-policy/

There's likely nothing your sales or marketing team can when Google, and companies that cargo cult Google's legal, decide not to allow your product gain a foot in the door because of its license.

Re: I wouldn't invest in open-source companies, even though I ran one

#169

Post by hodgesrm » Mon, Jan 25, 2021, 6:50 PM UTC

jandrewrogers wrote: post #135

hapless wrote: post #90

The author makes a good case against open source businesses, but he indulges himself in a false dichotomy between a traditional proprietary model and open source. Increasingly often, there is no market for your proprietary software. It is then not a choice between open source and proprietary models, but rather between doing open source or doing nothing. A lot of markets are just gone , and they're not coming back. Od…

The market for new proprietary databases is both large and lucrative -- it is printing money right now. What has changed is the structure of the market and the product economics. The classic marketing-driven retail database product model, the kind that puts billboards on the 101, stopped working as a financial model a decade ago -- development costs are too high for the marginal return. That has been replaced by mode…

> Proprietary databases have found their niche with qualitatively higher performance and throughput -- often 10-100x -- than the best open source has to offer on the same hardware footprint.

This is not true in general. It depends on the workload. Here's an example: MySQL 8.0 is benchmarked at over 1M QPS on I/O bound reads using Optane SSDs. [1] You would need to show that proprietary software hits 10-100M QPS for this workload and hardware configuration. Oracle and SQL Server are often faster because they have better query planning but it's not orders of magnitude.

[1] http://dimitrik.free.fr/blog/posts/mysql-performance-1m-iobo...

Re: I wouldn't invest in open-source companies, even though I ran one

#170

Post by Tomte » Mon, Jan 25, 2021, 7:10 PM UTC

Andrew_nenakhov wrote: post #32

xyzzy_plugh wrote: post #20

Earlier quoted context omitted.

This is a small point to argue. Even if you choose to distribute your source on-demand by CD in an envelope with postage stamps on it, the next logical step is for a customer to just put that code on github. So, what then is the difference? Putting it on github (or a zip file on your website) is less of a hassle.

If the customer paid a really high price for software and source code (like a few million dollars), he might rightfully decide that publishing it and giving it to everyone for free is not in their best interests.

That used to be the business model of Adacore and the GNAT compiler.

Their compiler was technically GPL, but they sold it for thousands of dollars to businesses, with the implied threat that they would let the phone ring for a while when a customer who distributed the sources had a problem.

Since their customers are using it for safety- and mission-critical software, noone would ever dare to cross them.