In a poly repo setup, agents are less effective having to infer changes across repo boundaries using specs rather than code as context. Changes that impact multiple repos are also much messier to wrangle.
Monorepos come with a lot of pain too. Two sides of the same coin. I manage the build system for a large monorepo. Questions that will get you to a primary source of pain...
How do you minimally build based on the changeset? How do you know this is sufficient for correctness? What happens when feature branches get out of date and don't see the upstream change that breaks the local branch? How do you version subprojects, as they change or as a whole?
Monorepos have a habit of creating hidden dependencies. The languages you use can help or hurt here.
Is there a combination of open standards to drop in to replace AD integration with self management?
OAuth enabled systems aren’t enough, central management of users and machines are huge. If that core matures, it opens up the market for replacements in other areas. Teams, Outlook and the Office Suite need first grade replacements.
The equivalent self contained home server exists today in the homelab community, either with Mac Minis or NAS systems running Unraid or TrueNAS with community apps. Add in Tailscale on top for remote access.
What’s needed is a lot of work on the software front to make it much easier, with interoperable standards. Self-hosted WYSYWIG options as easy to use as the social media tools for photos and writing and social posts. Ability to run distributed chatroom style instances with tracker like discoverability to replace Discord. Built in backup options with easy offsite backup replication.
It’s a 2 x 4 Lego brick with a speaker/lightsand custom ASIC built-in, with light and sound sensors, reacting to IOT beacons that allow different sounds or light sequences. And it’s rechargeable like an electric toothbrush. It also has accelerometers that change the sounds as you twist and turn them around. The sounds themselves are generated, with the speaker driven by an onboard synthesiser.
The most concrete part I get is that it has a recognition and positioning system built-in, so it can recognize the IDs and relative positions of nearby beacons. Beacons are inside bricks, tags and minifigures. The bricks seem to also have some kind of color sensor to detect the color of nearby normal bricks.
Then it does ... something with that information.
From the promo it almost looked as if that data was fed to an LLM that could then generate an audio response that fits to the play scene. Something like "You are a <Lego Star Wars minifigure>. You are <sitting> in a <vehicle: air plane>. The vehicle is <turning along the Z axis>. What do you say?" (Where the stuff inside the brackets is inferred from the nearby beacons and the rest would be a fixed prompt template)
But that would require the bricks to have an internet connection, and I have no idea if that's the case.
Yep. The initial marketing video makes it seem like a "brain+motor" super-brick that can somehow turn minifigs into robots and autonomously drive vehicles. But no, it's just a speaker + couple LEDs and a proximity sensor.
I’m using a similar format to keep my links organized on an iPhone. A bookmarks.txt note kept in Apple Notes with markdown # headings to separate each category, allowing me to not keep open 500 tabs on Safari.
Let me check... 1331 open tabs in Safari on my iPad. Several text files with thousands of bookmarks, with various attempts at grouping. Since I cannot save myself, I save them.
But as an Obsidian power-user, I regularly paste screenshots into notes
There is a plugin that allows templating the screenshot file name, so naming the pasted screenshot, using the same as the note where it's being pasted, and a timestamp, for example, is easy.
Open source licenses as they exist today aren’t sustainable to run a business. We’ve seen with the cloud providers how easy it is to launch a competitor if you don’t have protective licensing. Gumroad’s licensing is still small business friendly and protects another Gumroad clone from being launched.
I would argue it is possible to run a business and be sustainable on open source, it's just harder and is not so compatible with the growth that many want.
I don't have an issue with this kind of license being used where open source does not suit, but I don't think we should change/widen the definition of "open source" to suit the sustainability needs of those that open source isn't compatible with, at the impact of the freedoms and open rights it provides.
The problem is that if you're not already differentiably the best at hosting your service right when you launch, someone else that's better at hosting can just do it and take all your business.
And hosting while keeping your prices down is not just a whole different skill set, anyone that's already a big will have pricing deals with AWS so they will beat you even if you host in the exact same way.
It's probably less differentiable in the case of something like Gumroad which is less likely to have big scaling problems, but for things like a distributed database, you run a serious risk of someone who is paying AWS half of what you are per compute hour just deploy the Helm chart and undercut you completely.
I've been sustaining myself for a couple of years now on my open source project (BookStack). Still going in a positive direction.
Other than that, some that come to mind: Proxmox, Opnsense, SnipeIT, GitLab, Canonical, Codeweavers/wine, Plausible, Home-assistant/open-home-foundation/NabuCasa, FreeBSD Foundation, Laravel, Blender, Godot.
Within there is a whole mix of business plans, some offer hardware, some are open core, some offer related paid services, some offer hosting, some offer support etc...
Which is fine. Not everything needs to be defined to be suitable for businesses. It's even fine for things to be defined to be explicitly not suitable for a business.
> Gumroad’s licensing is still small business friendly and protects another Gumroad clone from being launched.
That's fine and dandy, but that doesn't inhibit me from rewriting the code from scratch and creating a clone myself by just matching Gumroad's existing feature matrix.
As a marketplace platform, it’s still lower than Apple/Google/Valve’s 30% cut. You pay for distribution, security, pre-integrations, shopping cart and other capabilities if you don’t want to do your own software development.
In a poly repo setup, agents are less effective having to infer changes across repo boundaries using specs rather than code as context. Changes that impact multiple repos are also much messier to wrangle.
reply