Yeah, it has a total lack of any of the integrations that made Assistant actually useful.
Even if an LLM offers meaningful improvements on the core conversational elements of Assistant, launching it without all the needed integrations is idiotic and completely hamstrings the functionality.
Notesnook is fucking fantastic.
I have spent over a decade - no I am not fucking joking I genuinely mean that - searching for a good Evernote / Onenote replacement. I have tried everything. Obsidian, Joplin, Silverbullet, Trilium, etc, etc, etc, etc, god I have forgotten the names of all the different note apps I’ve tried. They have all sucked. Joplin sucked about the least, but it still never really convinced me to get my stuff off of Onenote.
Notesnook blows them all away. Syncing is instantaneous (literally, you can type into a note on your phone and watch the words appear one at a time on your laptop), you’ve got S3 storage for attachments, sharable notes that can be password protect and set to self-destruct, lockable notes, read only notes, everything is exportable in multiple different formats, notes can be linked to multiple notebooks, notebooks can be nested, notes can be tagged, there’s bi-directional notebook linking, an attachment manager, every note has an auto-generated table of contents, the WYSIWYG editor is beautiful and works flawlessly, they have a web-app (unlike Joplin or any of the other commonly recommended solutions), there’s a web clipper that works really nicely with multiple different clipping formats, the phone app has one for one feature parity with desktop and web, they’ve got an absolutely beautiful code-block system with a copy button built right in so it’s incredible for storing config files or instructions for a self-hosting process… I could go on but I think I’ve ranted enough.
Also, just to be clear, Notesnook is fully self-hostable. There’s an excellent guide here: https://sh.itjust.works/post/31407921. If you self-host, you get all the pro features automatically.
You can host the web-app as well if you like - it doesn’t have a dockerized version yet, but the code is all up on their github - but you can also use the web-app on their server to connect to your back-end, so it’s really not necessary.
The documentation from Notesnook sucks, but this tutorial will have you sorted: https://sh.itjust.works/post/31407921
I’ve never had any luck finding anything open source / self hosted that actually works reliably. The best I came up with was an Nvidia Shield, which at least allows to you to run stuff like Kodi, Jellyfin, and Smart Tube Next (ad blocking YouTube client).
Sorry, you’re correct, that was a typo, I’ll edit.
The point of comparison is still correct, about it being equal to the total compute of London. I just wrote the number wrong.
If AI was actually going to replace every human worker on earth, even with huge efficiency gains you’d still want to build out absolutely ungodly amounts of compute capacity.
Also, it should be noted that while Deepseek has demonstrated that its possible to substantially reduce the compute requirements for transformer based models, doing so relies heavily on a “Good enough” approach that moves the results further away from being enterprise capable. It’s not a cut and dried solution to the backend costs of running these models at the scale that investors want to see them running.
That’s part of it, but the bigger part is that the enterprise business has completely failed to manifest.
Transformer model based AI is, at best, a fun toy or a minor convenience. For example, as a coding assistant, it functions as a fairly effective and speedy search engine, so long as you have the skill to check the output. But it also shortcuts the research process for you in a way that makes you more likely to overlook any potential downsides in the solution it offers, or perhaps miss a better solution you might have found if you did the research yourself. It’s handy for stuff that’s low stakes, but not reliable enough for anything that really matters.
And as a replacement for a coder, it just sucks, producing bug riddled, insecure code that it lacks the ability to debug effectively, meaning you still have to employ coders to fix its output.
This is the story with every potential entreprise application; it can’t be trusted enough to replace the expensive humans it’s supposed to replace, and it doesn’t actually make the expensive humans substantially more productive when it assists them (this has been studied; while a lot of people will anecdotally claim AI makes their job easier, in practice the numbers show that it mostly either slows people down or makes no difference).
As a customer service agent, it’s a dangerous liability, with a habit of outright hallucinating answers, and vulnerable to prompt injections that could allow for all sorts of dangerous shenanigans if you give it the ability to actually start making decisions about refunds and rebates.
Replacing troublesome, expensive humans was always the real sales pitch. Every other feature these charlatans advertised was just part of a scam to keep us troublesome, expensive humans from complaining too loudly about getting replaced. When Sam Altman tells us that AI is going to cure cancer and solve global warming, he knows it’s a lie. But he also knows that “I want to take all your jobs” isn’t exactly great PR, so he has to invent reasons why this is actually a good thing.
But the real sales pitch hasn’t come through, and it’s increasingly apparent that it won’t. You don’t invest hundreds of billions of dollars into compute to sell a slightly “better” search engine, or an app that shows you a short video of you kissing your crush. Those are not products that are going to give that kind of return on investment.
I’m not saying that transformer model AI will never have any practical uses. But the big practical use, the one that was driving a truly unprecedented wave of tech investment; that isn’t happening. AI is not going to come and take all our jobs. Or, at least, this version of it isn’t. Microsoft have made the smart play by jumping off the runaway train before it reaches the bridge. They’re going to get fucked by this, but not nearly so hard as the people who don’t know when to fold.
Microsoft hasn’t signed because unlike the others they’re scaling back their datacenter plans… Massively. As in, they’ve cancelled 1GW of future builds, which is equivalent to the entire compute power of London, AKA the datacenter capital of Europe.
MS has seen the writing on the wall; AI is a nothingburger. And keep in mind, MS basically owns OpenAI, the apparent leaders in the field, and has access to all their tech and IP. If Microsoft are calling bullshit, it’s not for a lack of information. They know exactly what Sam Altman is cooking up in secret, and it’s clear from their reaction that they know its just a new coat of paint on the same busted crap.
If OpenAI really were quietly working away on AGI, or some magical new version of ChatGPT that solves all its problems, MS would be in the best position to get out ahead and profit from that. They’d be building capacity and power generation like crazy.
This whole bubble is primed to burst.
If you’re looking for a good Nextcloud replacement I recommend Seafile. Been using it for years, very solid.
So, unfortunately, this latest update seems to have created a lot of issues. First off, MobaXTerm support appears to be borked. Second, attempting to connect directly to LXC containers throws an error because I haven’t linked a WSL2 instance for X11, even though X forwarding is not enabled for the connection.
I’d suggest maybe testing with a plain Debian or Fedora install. Just enable KVM and install virt-manager, and create the environment that way.
Unfortunately I’m not very familiar with Cloudstack or Proxmox; we’ve always worked with KVM using virt-manager and Cockpit.
Our usual method is to remove the default hard drive, reattach the qcow file as a SCSI device, and then we modify the SCSI controller that gets created to enable queuing. I’m sure at some point I should learn to do all this through the command line, but it’s never really been relevant to do so.
The relevant sections look like this in one our prod VMs:
<disk type='file' device='disk'>
<driver name='qemu' type='qcow2'/>
<source file='/var/lib/libvirt/images/XXX.qcow2' index='1'/>
<backingStore/>
<target dev='sdb' bus='scsi'/>
<alias name='scsi0-0-0-1'/>
<address type='drive' controller='0' bus='0' target='0' unit='1'/>
</disk>
<controller type='scsi' index='0' model='virtio-scsi'>
<driver queues='6'/>
<alias name='scsi0'/>
<address type='pci' domain='0x0000' bus='0x04' slot='0x00' function='0x0'/>
</controller>
The driver queues=‘X’ line is the part you have to add. The number should equal the number of cores assigned to the VM.
See the following for more on tuning KVM:
What are your disk settings for the KVM environments? We use KVM at work and found that the default configuration loses you a lot of performance on disk operations.
Switching from SATA to SCSI driver, and then enabling queues (set the number equal to your number of cores) dramatically speeds up all disk operations, large and small.
On mobile right now but I’ll try to add some links to the KVM docs later.
Been using Xpipe for probably over a year now. It’s amazing and I wholeheartedly recommend it.
Yeah, self hosting gets you all the pro stuff. And you can even use their web-app and point it at your backend server, as well as the desktop and phone apps.
The Notesnook mobile app is really slick.
For anyone who has been recommended Joplin and found it kind of clunky and ugly, I want to throw a shout out to Notesnook. You can self host it, it’s beautiful, and it’s absolutely brimming with great features. And any of your notes can easily be exported in markdown so there’s no lock in.
Here, https://github.com/streetwriters/notesnook-sync-server
But I recommend this thread as a better setup guide; https://lemmy.ml/post/25006407
You have to actually create your account on a self-hosted instance. Self-hosted accounts are all Pro by default.
I believe that’s only if you use their servers. You can selfhost.
No, it’s just a bad take. Every other manufacturer of self driving vehicles (even partial self driving, like automatic braking) uses LiDAR because it solves a whole host of problems like this. Only Tesla doesn’t, because Elon thinks he’s a big brain genius. There have been plenty of real world accidents with less cartoonish circumstances involving Teslas that also would have been avoided if they just had LiDAR sensors. Mark just chose an especially flashy way to illustrate the problem. Sometimes flashy is the best way to get a point across.