I'd normally agree, but Joshua (the founder) has been minimally involved for years outside of the patent lawsuits. With that mostly resolved (in Mycroft's favor), he's moving on to the next thing.
If you want to replace a founder sometimes we just push them aside to make the focus on a new qualified CEO to take the company to the next level. After awhile people get used to the founders not being highly involved and their departure becomes less of a shock.
Sure. I joined the company last year, and Joshua has mainly been busy appearing in court to fight the patent troll. With that fight effectively over, Joshua and Kris are focusing on something new (not sure what specifically).
Michael Lewis has been running the company for two years, so this resignation is more symbolic than anything.
WordPress, contrary to what most think, is pretty fast and performant. The reason why so many WordPress sites are slow and can't handle traffic, is because they are run by marketing people who don't understand the tech, and load it up with dozens of poorly written extensions to add user tracking, calls to action, and other dynamic behavior to increase retention. These sorts of practices have nothing to do with whether the framework itself is fast or not.
As has already been mentioned as well below, it is trivial to set up caching on a WordPress site. Most of the WordPress sites that I run are entirely served from cache by nginx. Some are just CMS APIs that I use for a next.js build, and others have the simply static plugin which will rebuild the assets whenever something changes. It's really quite an amazing system, and as much as I do not enjoy PHP or WordPress, there is a reason why it is so ubiquitous now.
WordPress does have a serious problem of underqualified people writing plugins for it and such, but again that is not WordPress's fault, unless you want to blame WordPress for being successful enough to attract those people.
Yeah but I mean: you can't really say that your homegrown Perl script running on a potato can handle 10k req/s if you put caching in front of it. What you're really saying in that case is that your caching solution can handle 10k req/s, which is fine, but besides the point.
Same with Wordpress. Wordpress itself cannot handle that amount of load, which is why you put caching in front of it. Users won't care, I give you that. But if you're talking about how much load a specific software can manage, including putting caching in front of it kind of defeats the point of the conversation.
There's a difference between full page caching like you're describing and in application caching. Most Wordpress caching extensions work internal to the application, not externally by putting something in front of it. With one of the popular extensions you can reduce database calls to basically zero without putting a http cache in front of the application.
I think the implicit point made by GP is to question what the purpose of database calls are to deliver a blog post. If the post were just some HTML and CSS, it should be able to be served with very little processing power. It's unlikely that dynamic template population is needed for blog posts.
> If the post were just some HTML and CSS, it should be able to be served with very little processing power.
Actually, with caching - they are. The database is there so that the person writing the blog post doesn't need to be very technical, or maybe so that they don't need to make a special effort just to update some post.
It doesn't make sense to talk about "how much traffic WordPress can handle" because WordPress is a software that runs on a server. It's like talking about how much traffic nginx "can handle" -- the server setup, hardware, RAM, etc. are key. WordPress running on a potato will handle less traffic than WordPress running on a decent server.
I replaced WP long ago, first with hugo (bad decision), then that with ghost (better decision). Not because of traffic, but because of the sheer number of hacks against the code base of WP appearing as zero days, and the rather extreme upgrading pain (the zero-touch upgrades failed on me a number of times, and I had daily DB backups). I had caching turned up to 11, and that system could easily handle the load (at the peak I was getting a hit every few seconds).
Hugo is/was spectacularly terrible from the perspective of authoring rich content. I could do as well creating a directory tree and writing html by hand. Or markdown.
I guess as I get older, I'm less inclined to spend time on things that don't just work, and don't get in your way with (often bad) opinions on how you should work. WP felt like too great a risk to work with. It was getting expensive to buy plugins to secure the site, when it wasn't generating revenue to offset the costs. Hugo was simply painful to work with. Ghost seems about right. Though grafting in the commenting bits took a little bit of work, which violated my concern about just working.
There are no great blogging solutions. There are a bunch of options of various capability. WP needs caching/php-fpm to work at an acceptable level for reasonable traffic. Hugo (as static pages) wound up not needing much caching. Ghost doesn't need much caching either.
It can be both, but you can have php-level caching (opcache or similar, mainly for making the actual php code faster), wordpress cache plugins (not that great at speed but they can invalidate their cache on changes), reverse-proxy level cache (varnish/nginx or similar, fast but usually requires some work to do invalidation) and perhaps a CDN cache (usually with the cache time set by wordpress via a cache-control header).
For cases like this at least having a nginx/varnish cache with a use use-stale-on-error policy would have been helpful. Even a very short cache-time like 1 second would probably have given a decent hitrate.
I've started using the Simply Static plugin a lot because you can get content-aware cache invalidation and also have nginx serve the files directly so you can avoid hitting any php. You do have to configure nginx properly though or else it won't work, but I didn't find it terribly difficult to do (although I'm already pretty familiar with nginx). Adding varnish is something on my list to try as well, though so far performance has been beyond "good enough" so it's hard to justify the time investment to adding varnish :-)
If you have to use a cache plugin on wordpress, it's much easier to use apache instead of nginx because all cache plugins can automatically configure the cache behavior using automatically generated .htaccess files instead of requiring you to manually configure nginx yourself. Nginx does have speed advantage over apache (and other advantages as well), but if majority of your traffics are going to get served by a static cache, might as well use apache because it's not particularly slower at serving static files compared to nginx, while saving you some configuration headaches.
You're acting like setting it up by hand means you're diving into the code and doing it yourself. In truth it means just picking the caching extension that matches your needs the most, and if you don't know what that means picking the most popular one will be just fine.
Personally I like this. It means that I was able to pick a caching engine that integrates directly into nginx, so that when I write a new post or edit an existing one the extension invalidates it directly in nginx. This means for the vast majority of my site the content is being served directly from nginx's memory cache. My performance is awesome. If they didn't go with an extension approach that would not have been nearly as easy to accomplish.
Also, because the site is overloaded, the non-archive link opened up with the browser tab title "We've had some trouble," which puts a different spin on the whole thing.
That's not an inference you should draw. I've known many people who have serially started companies. Starting a new company doesn't mean that the earlier companies failed or that the founder stayed at those earlier companies instead of selling them or hiring someone to manage them.
In the case of the CEO it probably refers to the decade he spent building and then selling two companies for around $200m, the larger one of which he founded.
It usually takes at three generations of highly-successful individuals:
Gen 1: Manual labor, kid goes to college and gets a professional job
Gen 2: Professional has a decent home, can provide a basic safety net, and can push / encourage / raise kid to be academically competitive at an elite level.
Gen 3: Starts a business which reaches millions of dollars / enough to be independently wealthy.
You can add one more generation if you also want to break into the monied power class. That's on the very, very fast end. It's usually much more than that, since it's rare each generation will have a highly-successful individual.
It's better than no mobility. We had no mobility just a few hundred years ago.
It takes many forms afterwards. I think many here/on this topic view this as a happily ever after scenario where then there is a family tree full of successful people doing interesting things at low personal risks.
My grandpa was a Gen 4 in your lineup. He was essentially a trust fund kid that never worked and just squandered the significant savings and assets of his depression era parents. My dad got some rather minor assets. Or so I heard growing up. In reality he got a free house, the one I grew up in, and several debt free rental units/cash flow. Not luxury apartments but growing college town type market. He was essentially a minor league trust fund kid who was bitter about having to work to afford everything he wanted (something he never saw his dad do). He kept a mediocre job and never strived for much. He was more interested in exotic cars, fashion, women, etc. He ultimately blew through all his inheritance in by my young adult years. I got nothing but a stable home environment growing up. Paid for my own college, cars, and everything else. I know many others in similar situations where gen 4-6 seems to be highly correlated to addiction, crime, etc. and just never strive for more.
I feel like I’m just starting over with no real advantage. Similar to maybe what gen 2 had.
This is my translation from "富不过三代”, an old Chinese saying about the impossibility of sustaining wealth.
In some similar way, the power and wealth of a nation also is unsustainable. Every powerful nation falls eventually.
And even the whole earth undergo patterns of up and down. 20th century seems a peak of human civilization, and 21 century seems a slow match to the inevitable crash...
Which, frankly speaking, might not seem fair, but is not necessarily bad. Over the centuries, people were doing things that were not directly useful to the society (like art) only thanks to being sponsored by someone else (who cared - in this case, parents). And even if after 10 years you fail to produce a startup than you can sell,* you are still left with tons of experience that you can use (yes, negative experience counts, too).
*) Or bring a positive net result, but this is a bit more difficult.
Does Mycroft still actually do something? So many years later it's still beta quality. And a dependency on Google is a total non-runner for me in terms of privacy.
I remember with the Mark II Kickstarter in 2017 they were kind of the Messiah of privacy-conscious voice control. I really wanted to back them but I didn't have the cash at time. Though that turned out to be great luck.
Because then they screwed that up and nothing happened for years. Now they continue but the project is more like a smart screen than a smart speaker now.
And all this time it's still not possible to run the core locally for true privacy. As it is it's just the same as Google, Amazon and Apple just with better promises on privacy.
Maybe I'm overlooking something because they have long fallen off my radar. But I'm still disappointed about the Kickstarter fail. They have a lot of trust to rebuild, at least with me.
At this point kickstarter campaigns should be considered donations with a low value and low probability of winning lottery ticket included with each donation. The expectation that you get anything in return from backing a kickstarter needs to be reset.
It's just Sturgeon's law/the Pareto Principle. There are good crowdfunded products that succeed, don't focus on the 90% that are crap.
> The expectation that you get anything in return from backing a kickstarter needs to be reset.
This is why Kickstarter repeats time and time again that it's not a marketplace. If you go into it thinking you're buying a product, you're gonna have a bad time. That's not at all what it is.
I've been a supporter of mycroft since the early days, even did some early work on integrations, but like most hardware kickstarters they ran into /a lot/ of delays (I think it was 2 years post projected date that I actually got my dev kit - but they did come through) and since then progress has been s l o w and it seems they've lost some of the value of an excited community.
I'd still love for an open source voice assistant to exist - especially since Sonos bought Snips a few years ago and doesn't seem to have done much with it, but not as excited about this project I was a few years ago
Nominally related: shout out to the awesome Mycroft search engine plugin project.
I wish it got more attention. I never understood why Mozilla built this half-assed search plugin feature on the addons site instead of integrating with Mycroft when it had already existed for years.
Joshua Montgomery founded Mycroft AI, but Michael Lewis has been running the company for 2 years as CEO. Joshua recently "resigned", though his day to day involvement has been minimal.
What's HNsers expo yet experience using Mycroft? I tried to play with it on a rpi a few years ago. Got stuck blue being able to use a Bluetooth mic. I found the CMU Sphinx voice was good to just use on it's own.
Open source voice assistant, not unlike google home or amazon alexa. The voice assistant itself works pretty alright but currently relies on google's speech-to-text to get good results. The technical infrastructure is a bit baroque and it's missing things you'd probably consider to be pretty essential, like an android app.
Their goal is to become the platform for embedding voice assistants in things like cars.
