jdvh

I got a Mazda for this reason. Physical buttons for everything and carplay for navigation+spotify. Best of both worlds. Infotainment has touch screen when stationary, but while driving it can only be operated by voice and console buttons/wheel.

Hydroelectric is the worst for catastrophic failure. See https://en.m.wikipedia.org/wiki/1975_Banqiao_Dam_failure

Nuclear reactors are extremely safe by comparison.

It’s by no means a given that wind/solar will be a major source of energy for the long term. If we’re lucky advancements in wind/solar and energy storage will make it a big part of our future energy supply. If we’re unlucky we’ve spent many trillions and made very little headway towards a more sustainable future.

Maybe the solution will turn out to be molten salt nuclear along with synthetic oil and carbon capture. Maybe something else altogether. Who knows, it’s still early.

I don’t think we should presume solar/wind will someday take over, and I’d like to see an “all of the above” energy strategy instead.

This is slightly offtopic but windows 95 ran on 4mb of ram. 8mb was considered luxurious at the time. Visual Studio 6, launched in 1998, ran fine with 16mb of ram. It had almost all of the debugging features we have today, and it worked flawlessly every time.

Half a gigabyte of ram was, and still is, a gargantuan amount of memory.

Getting harsh feedback isn't so bad, really. You get that kind of feedback either because people think your MVP is something it's not or because of quality issues, and you can fix both.

A bad landing page filters out people who are just moderately interested. They'll just close the tab and move on. That's going to skew the feedback you'll get.

If you don't get (much) feedback at all you're in a much worse spot. Then you don't know if your MVP sucks, or it doesn't but your landing page sucks, or maybe both are fine but you're just not getting the right traffic. It's way harder to figure out what to do when nobody seems to care at all about what you've built.

I think the core issue is that:

- Calendars are good at appointments but bad at everything else.

- Day planners are good for jotting down what you're working on and for revisiting what you did last week/month, but if you put anything with a deadline in your day planner you're bound to forget it. And if you have a ton of notes that relate to a single project, where do you put them?

- To-do lists tend to accumulate tasks and then you declare TODO bankruptcy, delete everything, and start over.

I'm convinced it's possible to merge these concepts intelligently, in a way that hasn't been done before. An app that has a free-form day planner, but you can still tag items with a due date, tags, etc so you can still see which deadlines are approaching.

You want to be able to zoom-in and zoom-out in your "tree" so you can do high level planning but you also want to be able to indent and add as many notes/ideas as you want without losing track of the bigger picture. With real-time multiuser editing and end-to-end encryption. That's what we're building with https://www.thymer.com

We're working on an outline app with task management/planning functionality. With end-to-end encryption and multiuser/teams support (you can see other users type like Google Docs). Early beta signup at https://thymer.com

Poorly insulated houses are windy with the windows closed because cold air drops. Air close to the windows gets cold (a single pane of glass doesn’t provide much insulation) and the cold air drops to the floor displacing the other air in the room, thereby creating a current. It’s like opening the fridge with bare feet. You can feel the cold air falling on your toes.

I agree that companies have a responsibility to listen to and address customer complaints. The problem is that a small percentage of customers create the bulk of the customer support burden. With phone support this gets even more skewed. You can't have engineers do phone support or they won't be able to get any real work done with all the interruptions. You can have engineers investigate and address all complaints that arrive by email. I strongly believe email support that goes directly to the people most knowledgable to fix the underlying problem is the best approach.

Companies that provide bad customer support just don't care. They'll provide phone support if they feel they have to, but they'll outsource it for cost reasons. Talking to people who are friendly but clueless is at least as frustrating for me as creating a ticket that doesn't go anywhere.

The cloud is so much more expensive than dedicated that you need far greater than 2x spikes in usage to make it worthwhile. For instance a video game that suddenly goes viral and you need to scale up 1000x overnight.

All cloud providers have pretty poor uptime records. Unless you set up multiple geo zones (extra complexity) your app will go down when aws-east does.

Solving self-driving in a select group of cities with permanent 5g connectivity, hi resolution gps, and 30k worth of sensors is a massively easier problem than solving “Tesla fsd”.

For this bet cars can deal with unusual circumstances in all sorts of suboptimal but harmless ways. Pulling over, stopping, going very slowly and cautiously is all allowed. You’ll still get to your destination safely.

I’m with Carmack on this one.

Plain text is so compelling because it's as simple as it gets, you can bring your own editor, you own your own data, and you can use version control.

Text+ is compelling because you can have images and some kind of formatting. You want to store metadata and have backlinks and tags. Ideally with the possibility of collaborative editing.

There should be a way to fuse these two.

What's the attack surface of a git linux account that doesn't have a shell and can only log on through ssh with a key from whitelisted IPs in a firewall? Compare that attack surface with that of a complicated web app like gitlab.

This service update only applies to people who self-host. Servers managed by gitlab are patched by gitlab staff. If you're going to self-host I think you should at least --consider-- self-hosting something with a tiny attack surface instead of a complicated product with many dependencies and "casual" security practices.

Either go with a hosted solution or take the time to learn linux server admin. Linux hasn't changed much in the past 15 years, and I don't think it will change much in the coming years either. With web stuff you have to keep up with new developments, but linux is very stable and boring by comparison. There's a learning curve, but taking the time to understand how linux really works is totally worth it.

You'll want to follow some linux hardening guides, set up a firewall, set up fail2ban, set up full disk encryption, backups, and some other things.

It takes about 30 minutes to set up your own git server. You can easily create your own git hooks to send email notifications on "git push", or to trigger CI or deployment. I get that gitlab is tremendously valuable for bigger teams, but smaller groups can easily do without.

This isn't the first time gitlab has security issues, and it won't be the last. It's just not worth it for us.

https://git-scm.com/book/en/v2/Git-on-the-Server-Setting-Up-...

I was a semi regular on BoS maybe 15 years ago. It was often interesting, and I remember the focus on desktop apps. Tons of SaaS skepticism.

We're building a todo/planning app https://thymer.com. It will have a free version for individual users (and paid for teams).

Although it doesn't have a kanban board (at this stage), our goal is of course to offer a good solution for everything related to planning and scheduling.

We're strongly considering e2e encryption, but we're not 100% sure yet, because other features like CALDAV integration and server-side search are not compatible with e2e. The app runs entirely locally and syncs in the background (via CRDT), so it's super responsive.

Tables can be arbitrarily complex, just like html, so yes a line has to be drawn somewhere. Spreadsheets and slide presentations can be copy-pasted as file attachments. That's fine.

You can't have a table look the same in every app, nor is that desirable (you'd have to embed fonts and re-create postscript). What you can have is the ability to copy-paste a table (or a bulleted list) back and forth between apps without mangling the state in the process. Rich Text intertwines semantics and style and it's not a good solution.

Most of the time you don't need the layout precision of postscript. There is a huge difference between not having any kind of tables or semantic context for text and having some that cover the 95% of daily use.

We have a binary and universal format for plain text (utf8) that is extended all the time with additional emojis. But it seems almost unimaginable that we ever get a binary text format that also allows simple semantic markup, like emphasis, tables, hyperlinks, amounts, dates, and phone numbers.

It's 2022, and really, we should be able to copy a paragraph from an email with a table in it into a chat app and have it just work. HTML is way too complex to use as data communication format between apps (vulnerabilities; every app allows a different subset). Markdown is made for humans to read/write, when what we need is a simple unambiguous binary format that is easy to parse.

We can have compound emojis where a polar bear is a bear + a snowflake joined by a zero-width joiner. But a datetime or number in text is too much to ask?

My approach is really pragmatic. You're unlikely to mix enums that have nothing to do with each other. The only issue in those enums that you could possibly get mixed up. For those you can just use non-overlapping integers for the values. It also makes debugging easier "over the wire", because you're probably not encoding the enums as strings. There is really no reason to have all enums map to the same handful of integers. If you don't do that, it just works.

Your competitor may have crazy momentum today, but who knows about the future. Companies with bright futures blow themselves up all the time.

Startups with momentum that don’t self-destruct are likely to get acquired. VCs happy, founders happy. Senior engineers will bail soon afterward and the mood will change. More unhappy people on twitter. Bugs linger and don’t get fixed. The business is now focused primarily on getting larger customers or expanding to different markets.

And that leaves you with plenty of opportunity. It might not look like it now, but their 15 minutes won’t last forever. Be happy for their success and be happy that they’re expanding the size of the market for everybody. Keep plugging away and focus on your customers.

Kudos to those who launch. It’s always harder than you’d think. That said, I don’t understand why people go the gimmicky route the moment they want to make something commercial.

Even a micro saas has to provide something of real value. You can’t skip that part, and surprisingly, it’s much easier to charge money if your initial focus is just on making a really good product.

Even encryption algorithms proven correct later turned out to be critically flawed.

The legal system is such a quagmire because figuring out what is "right" in all scenarios is anything but obvious. That's how we end up with laws we don't enforce and technical legal behavior that will still get you into trouble. But what's the alternative? Robot justices? No thanks.

If the complexity is in the nature of the problem, then smart contracts must contain all of this complexity. But how can you ever write smart contracts that are bug free and deal with every contingency?

Instead of making systems that are buggy, cold, and unforgiving we should be making systems that are more tolerant of human mistakes.

The scenario you describe hasn't happened to us. We haven't had more than a couple of minutes of unscheduled downtime in the last 10 years.

That's fair, and I'm not dogmatically opposed to the cloud. But when you have underpowered cloud machines you end up having different machines for email, for web workers, for caching, and so on. The complexity you introduce by making your app scalable in that way you just don't have when you stick everything on 1 dedicated server. With dedicated machines I don't need to create a distributed architecture prematurely.

It takes us some time to set up dedicated machines, but less than a day. Ansible helps a bunch.

We've built a couple of saas businesses without any investment, as complete unknowns. It's doable, and today it's easier than it has ever been. You can go from 0 to revenue in less than 6 months. The big problem is that there is a ton of software out there, including free software, and people will only pay if your product is much better in some way that actually matters. If you only give yourself a few months to build a product you can only build a couple of features, so those features better be really good.

Hi! Today we're launching a full rewrite of our wiki product Papyrs. We're particularly proud of our new editor that -- in our view -- achieves an impossible blend of markdown and rich text editing that appeals to both power users and occasional users. A pure drag&drop rich text editor is great for demos, but when you create a ton of content having to drag widgets gets old fast. Markdown is great, but our product is meant for businesses and pushing markdown on nontechnical users doesn't work. And a lot of rich content like calendars, forms and image galleries aren't suitable for markdown at all. So Papyrs does both, and users can simply switch between rich text and markdown editing modes as desired, and it's kind of magic!

If you want to play around with it you can go to https://papyrs.com/try/ and get your own playground instantly.

All feedback appreciated!

> It's analogous to IQ tests having cultural biases, but people's claiming they measure intelligence objectively.

No contradiction there. IQ tests clearly aren't completely unbiased and perfectly objective, but they don't have to live up to that impossible standard. Humans have a general intelligence factor G, and an IQ test can approximate it reasonably well. As far as I know IQ tests are the best and most objective way we have of measuring G.

If we can't even call our most objective tools objective, then we're quickly entering the territory of total relativism.

The idea that the outcome of IQ tests is mostly determined by cultural biases, which is what you suggest, seems extremely unlikely. Cultural aspects will make some demographics score better than other demographics, but smart people within a demographic will surely still outperform their peers. So I'm very skeptical about your claims.

From what I remember from The Paypal Wars (highly recommended):

There was a culture clash when X got bought by Paypal. Elon was in charge of X and became the CEO of the combined company (as Peter Thiel took a leave of abscence) and wanted to switch from Unix to Windows. Max Levchin, the main tech guy at Paypal hated that idea. Elon initially got his way, and Max got marginalized. He then started looking into fraud cases and discovered that fraud, although low as a percentage of revenue, was growing rapidly. Max then persuaded people that the #1 priority should be to fix the fraud issue before it would kill them, and that they didn't have the time to worry about the technology stack. Elon Musk got kicked out, Peter Thiel came back to Paypal as CEO and Max managed to squash the fraud problem. Their competitors either failed to get traction or got crushed by fraud. With Paypal as the last one standing they won and got bought by Ebay.

http://www.amazon.com/The-PayPal-Wars-Battles-Planet/dp/0974...