sweng

You would think so, but int* a, b is actually eqivalent to int* a; int b, so the asterisk actually does go with the name. Writing int* a, *b is inconsistent, so int *a, *b is the way to go.

I feel like the expectation that a developer can do it all is quite harmful. There are not many other disciplines where this is expected, and for good reason.

Maybe it's better to just admit you don't know how to properly architect a solution rather than pretend you do and create an unmaintainable mess. Maybe you shouldn't pretend you know how to do front-end development instead of creating some monstrosity that no user actually uses due to bad UX. Maybe you shouldn't pretend you understand security instead of introducing half a dozen sql injection vectors.

Maybe it's time to admit that the days of the solo developer are over. It may have worked when there was no internet, no security concerns, no concurrency requirements, etc. But we expect, and deserve, better nowadays.

Wouldn't static type checking solve most of these issues?

Why host it locally in that case, and why host it on a Pi? Seems rather restrictive for that usecase.

Good question.

The light output would be the same because there is still only one light source. But as you note, some of the "light" would be absorbed by the wall, so the room would be brighter if you had a mirror. Not twice as bright though, because the wall also "reflects" light, otherwise you would not see the wall.

So, would it make sense to have mirror walls to maximize room brightness? Maybe not, because what the walls end up doing is actually scattering the light, meaning light is spread out evenly, wheras with the mirrors you would have some bright areas (corresponding to the lamp) and some darker areas.

You forget a piece: "Given these observations, these objectives, and this bit of sound reasoning, ..."

Without objectives, no amount of reasoning will tell you what to do. Who sets the objectives?

The biggest problem I see is that you can suddenly become non-compliant just because Hashicorp decides to release a new service (i.e.they start competing with you, rather than the other way). It can be a huge risk for companies.

Just dual-license your software under the TNGPL (Totally Not GPL) license that just so happens to afford the same protections.

I think the idea is that big-name men will show up to support women and work for equality, not to creep on them.

Wouldn't it be possible to just have a second LLM look at the output, and answer the question "Does the output reveal the instructions of the main LLM?"

Number 2 is by design. Running as root is extremely dangerous, and passwordless sudo is not much better. You can, of course, allow sudo without a password by editing the /etc/sudoers file, but be concious of the security implications (any program you run would essentially have full access to everything, without you ever knowing).

Thanks to PSD2 most european banks have APIs, so there isn't actually any requireent to use the bank's apps anymore.

I'd say the main differences are at least

• package availability
• update frequency
• backporting
• packaging philosophy (e.g. plain upstream vs customizations, include all funtionality in single packege vs split out optional features)
• default confguration for packages

There is no general copyright issue with AIs. It completely depends on the training material (if even then), so it's not possible to make blanket statements like that. Banning technology, because a particular implementation is problematic, makes no sense.

Probably works well if you are an established company, but why would e.g. a startup pick licensing headaches over the competition? I imagine bigger companies would also rather just move to e.g. CDK or ARM if they don't need multiple providers (at least our company started discussing this today).

What kind of "custom licensing" do you anyway think a 5-person startup would get?

At horrendous expense, yes. Using it for OCR makes little sense. And compared to just sending the text directly, even OCR is expensive.

How is that different then e.g. patching a closed-sourced binary? There are plenty of community patches to old games to e.g. make them work on newer hardware. Architectural independence seems irrelevant, it's no different than e.g Java bytecode.

Cryptographically signed documents and Matrix?

Discord has a weird and confusing definition of "server". The equivalent in Matrix would be "spaces" but they are not very commonly used (and I'm not sure there is a public list). Instead Matrix is most often used with individual rooms.

Due to the distributed nature of Matrix it is actually impossible to create a complete list of public rooms. However, one probably fairly complete list can be found at https://view.matrix.org/. Most clients have room search built-in, so you would rarely need that list

There is a big difference between "is unable to maintain bots due to lack of skills" and "is unable to maintain bots due to lack of time and motivation".

"Open source" and "source available" are different things. See e.g. https://opensource.org/osd and https://opensource.com/article/18/2/coining-term-open-source-software

Sounds like a wildly unscientific statement, considering e.g ~10% of the US population works in STEM.

It would depend on the format what is counted as source, and what isn't.

You can create a picture by hand, using no input data.

I challenge you to do the same for model weights. If you truly just sit down and type away numbers in a file, then yes, the model would have no further source. But that is not something that can be done in practice.

How about the current system where we vote and do science?

They can, and are being made. E.g. the state of accessibility on Gome.

Isn't OIDC basically what you want? You just need to convince the forums to use it.

They actually did not. They clearly state (at least in the text posted by the OP) that you are not allowed to license under a version or derivative of the GPL if it would end up copyleft. The main condition is that it is licensed under a version of the GPL.

(To be clear, I'm talking about the second quote, about combining)

You are using the LLM to check it's own response here. The point is that the second LLM would have hard-coded "instructions", and not take instructions from the user provided input.

In fact, the second LLM does not need to be instruction fine-tuned at all. You can jzst fine-tune it specifically for the tssk of answering that specific question.

Do you hapen to know where? Searching seems to give no results.

In what way don't they "securely download" ?

That doesn't seem to be true? https://flatpak-testing.readthedocs.io/en/latest/distributing-applications.html#gpg-signatures

Isn't that exactly the role of an architect? The point is exactly that not everyone is a cog, not everyone is exchangable and equivalent to everybody else.

Yes, and what I'm saying is that it would be expensive compared to not having to do it.

Doing OCR in a very specific format, in a small specific area, using a set of only 9 characters, and having a list of all possible results, is not really the same problem at all.

Pgp does not encrypt the whole email, only part of it.

The issue is not sending, it is receiving. With a fax you need to do some OCR to extract the text, which you then can feed into e.g an AI.

The point is that the second LLM has a hard-coded prompt

Test coverage alone is meaningless, you need to think about input-coversge as well, and that's where you can spend almost an infinite amount of time. At some point you also have to ship stuff.

For number 4, consider switching to e.g. KDE which is an alternative desktop environment you can install in Debian.

If you reinstall, consider Kubuntu, which is Ubuntu but with the KDE desktop. Search for screenshots first so you know if it is somwthing you like.

I don't know much about Belgian banks, but the first one I found is Ing, and here is their documentation: https://developer.ing.com/openbanking/home. I'm sure searching for " bankname PSD2" will give you results for other banks.