Metus

There is also https://open-assistant.io by LAION.

By supporting, participating in and using projects like this by LAION:

https://open-assistant.io

> The thing that is often de-emphasised in the presentation of the problem, in order to make it seem more mysteriously paradoxical, is that the presenter knows where the car is and this knowledge is always used perfectly. If the question always ended with "remember: Monty knows where the car is and will use this information", it would be more obvious.

The associated line of reasoning resolved the paradox for me. If I stick with my original choice, it is as if I ignored the new information. If I switch the choice, I react to the new information.

The extreme of this is to pick among countably infinite doors, having the presenter open countably infinite doors and leaving just yours and another closed. Who could reasonable suggest that the chance is still 50/50, assuming you don't flip a coin and base your choice on that?

Is there a fork running it on Apple's Neural Engine?

Can I somehow keep track of the content I generate? That is, the prompts, the answers as user and the answers as assistant. I only see my recent messages.

> Honestly I don't understand why a consortium of governments and businesses with high regulatory requirements don't simply get together and develop a common platform for this. They'd rather give billions of dollars to Documentum or Oracle. If they want support, SOMEONE will provide paid support, like Postgres

Or have a cooperative of businesses write the necessary software.

In Germany we have vaguely similar thing going on with DATEV: Basically all tax advisors are members of the DATEV cooperative if they want to use their software suite, which for all intents and purposes is able to implement absolutely anything a tax advisor is required to do while at the same time implement all regulatory requirements such as confidentiality, archival rules, reporting, logging etc.

In my opinion there should be a similar thing going on for all industry to implement the regulations required by GoBD, GDPR and so on.

Another general purpose backend could be S3-type object storage.

If I could store all kinds of documents there, I'd adopt in a heartbeat. My use case would be to combine this with a contract management system and to attach all (email) correspondence to the respective contract, so we wouldn't have to rely on the responsible people managing their inboxes.

In turn, I'd like to attach deletion, visibility and archival rules, such that I conform with GoBD and GDPR (I am based in Germany) where on one hand there are archival rules, such as keeping contracts for ten years after they are canceled, or correspondence for six years, but also have to keep in mind that PII gets deleted regularly.

So after a contract has been canceled I'd like to archive all correspondence and the contract itself, such that the operating team only sees active contracts and correspondence older than six years is immediately deleted.

This should be way higher up.

I find OpenAI having exclusive access to that kind of high-quality data more concerning than them having access to their current amount of compute and currently trained model. A couple of million dollars worth of compute is in the realm of any medium sized research university, larger company or any country worth of mention. And seeing as Moore's law still applies to GPU, the cost will only fall.

However high-quality data is scarce. I would be willing to fund a proper effort to create high-quality data.

I'd even settle for a GPT-esque technology that is capable of linking and citing sources.

I think an approach like this could generate interesting sounds we as humans would never think of. Or meshing two sounds in ways we could barely imagine or implement.

But of course something like this, which only thinks in 5s clips can not generate a larger structure, like even a simple song. Maybe another algorithm could seed the notes and an algorithm like this generates the sounds via img2img.

Good to know. I was just so close with just under 7s using 40 steps and Euler a as sampler.

> composition's lacking, too. It's loop-y.

Well no wonder, it has absolutely no concept of composition beyond a single 5s loop, if I understand correctly.

> It absolutely sucks at cymbals, though. Everything sounds like realaudio :)

> It could probably do bad modern production fairly well even now :) exaggeration, but not much, when stuff is really overproduced it starts to get way more indistinct, and this can do indistinct. It's realaudio grade, it needs to be more like 128kbps mp3 grade.

I haven't sat down yet to calculate it, but is the output of SD at 512*512px at 24bit enough to generate audio CD quality in theory?

Can you run this on any hardware already capable of running SD 1.5? I am downloading the model right now, might play with this later.

Guessing at the speed with which AI is developing these days someone is going to have the extension up in two hours at most.

And the Swiss ask for 0.13% and 1.01% depending on the Canton, pension funds are exempt.

The data for the website I work on shows four out of five people on mobile and one out of five on desktop. So I certainly would take your bet.

Could you use something like confidential computing to attest over http(s) what software is actually running on the server? This would offer a very interesting trust model together with reproducible builds, where you could have the CPU attest over http(s) that it is indeed the code base published on Github/Gitlab that is actually running on the server and receiving your data.

Could you expand on the benefits of double-entry bookkeeping, beyond reducing errors?

Personally I am looking into it, because the run-of-the-mill personal finance software applications can't even deal with the fact that paying off a credit card statement is not income and expense, but a mere change between assets and liabilities. Or buying some stock is not an expense, but a change between two types of assets. And so on.

However, the run-of-the-mill software is a bit better with categorizing things and integrating with other applications. Currently, I am looking for

- Connecting PDFs or images as invoices with transactions

- Marking transactions as relevant for my tax income statement

- Estimating my tax return in advance

and some more I can't think of off the top of my head. What weirds me out, is that everyone has to deal with personal finances, yet there isn't a single personal finance app that deals with everyone's situation. Do I massively underestimate the complexity of the problem domain?

After researching this for a while, it seems there is no documented, native option to do this. The only option is to unlock all SSH keys all the time, which makes them less secure than the passwords for websites managed by the exact same keychain. Which, in my opinion, is weird.

Do they employees at Apple use a different system altogether? Because the built-in one doesn't seem very secure. Or maybe I am using it wrong, who knows.

In a similar vein, is there an exhaustive manual for macOS? It bugs me that Apple machines cost a small fortune, the OS is full of nifty features, but there is no non-superficial manual shipped with it.

You aren't. It is the same with Yubikeys, if you use them. You are supposed to generate a certificate per user and device they are using for authentication, which is not a big deal at least for remote login on servers, as you can set an arbitrary number of valid SSH keys.

For anyone not aware, you can use macOS's keychain to store ssh key passwords and have them unlock at login. This way you can have the benefits and convenience of password managers in the command line for SSH certificates.

https://apple.stackexchange.com/questions/48502/how-can-i-pe...

Does it only work in the US or is there an offering for EU/Germany as well? Ordering pants online is a nightmare, as retailers do not know what length and width are.

What surprises me is that Apple does not make a distinction between commercial developers and open source developers, or any others that make only free apps. Both need to pay almost 100 bucks a year just to have the opportunity of publishing on the app store.

It seems to me that in the vein of 'commoditize your complements', developers that only publish zero cost apps shouldn't pay an annual fee for the privilege of making Apple's ecosystem more valuable.

Can you link to other examples of no code apps? This seems kind of interesting.

Related are the atrocious user manuals and tutorials for various consumer electronics products. You get descriptions like "if the self-diagnostic tool shows the message 'no errors found', it means that there are no errors that the self-diagnostic can see." Or other manuals or help texts that just plain rephrase the prompt.

Good technical writing is much like teaching a great course: Empowering the reader to use the product or service for their own purposes.

Can you link a document discussing this phenomenon or at least give the relevant keyword to search myself? I have noticed something similar when TA'ing, observing my colleagues and also when learning myself.

I really dislike that the saying is "money doesn't buy happiness." In German it is much more sensible: "Money alone doesn't make you happy." Which emphasizes the role of money as part of the formula.

Regression to the mean may be part of it. Even for a master of their craft success and quality will depend on random factors beyond skill, so pieces of work following a great one are more likely to be worse than better, assuming skill remains constant.

What does that setting do? There is no tooltip.

Thank you for the answer. So especially for personal use I am better off hoarding the ZIM version, especially considering there is the dedicated Kiwix reader, while I am not aware of a similar tool for the XML dumps.