Since there's a lot of talk about the CEF update, does anyone know why they compiled it in instead of having it as a normal plugin? It's incredibly annoying that compiling OBS means compliling almost all of chromium...
This should not be the case, at most you'd need to build the CEF wrapper, but the OBS team provides precompiled CEF builds for all platforms [0] or you can use the builds provided for free by Spotify [1].
Honestly, this is a clever way of doing this. It means the additional instructions are stored alongside the text they reference, which makes it easier for doc writers to keep in sync and it probably also makes it more effective since the bot is guaranteed to read it at the right time in the right context.
I doubt they actually send it with every request and hide it in CSS, it probably does some other kind of check (IP, user agent, token...). This also tracks with the fact it is now entirely gone - presumably they just fixed that if statement and it's again chatbot-only.
What? This is the most logical solution and the most likely one given the behaviour seen in the post. And even if they were sending every human reader the LLM content too, who cares? It's a few kilobytes at worst in a multi-megabyte page.
It still doesn't make sense. That pages are probably built from an internal repository/database where that information is in a human readable format (e.g. Markdown or some sort of CMS) that is then processed by a CI to produce the public static HTML pages.
It would be more efficient to give the chatbot context the source files, in a format that is more efficient for that (Markdown documents, for example) and that will of course contain the internal notes, that are stripped during the build of the public HTML version.
Why a chatbot should waste resources in parsing a public webpage for the same information that you have internally in possibly a more appropriate format?
Also having a check for user agent/IP to do so is still inefficient, because you can't serve static HTML/CSS/JS and thus you have to do processing work on each request, a thing that doesn't make a lot of sense for static content of this kind.
Nobody said it's efficient, welcome to the modern IT landscape. The docs are surely some SaaS with a WYSIWYG editor, access controls and analytics. Every page probably does a dozen database lookups and involves 3 mixroservices and 4 external providers minimum.
Given they probably need to go through layers of management, across teams, and go through relevant testing to adjust the LLM prompt, this was probably a nice quick and flexible way to create a "shadow" prompt.
Maybe? But it also doesn't necessarily belong into the prompt. The model probably accesses the help pages as a RAG knowledgebase, so putting it there makes sense.
There has always been the possibility that people would do dumb things with email they receive, but it wasn't a widespread thing to do, there weren't many reasons to do them and the tools weren't easily accessible to most people.
But LLM agents are really useful and the tools to give one access to your entire digital life is just an OAuth prompt away.
Soo many more things than banking apps at this point. Things as normal as looking at train schedules, grocery store catalogues, ordering a taxi... sometimes have SafetyNet DRM.
When you publish apps on Google Play, the Play Developer Console even nags you to use harder obfuscation and implement Google's DRM to, allegedly, protect yourself and your users.
Can the EU finally make this shit illegal? Oh wait, they're pusing for mandatory DRM too... Fuck.
So you can't pay a janitor enough to deal with vandals, but somehow the minimum wage bartender that already has to tend to a bar and cash register is able to keep their toilets clean no problem?
It's not that hard to make toilets that are nearly indestructible and can be very efficiently cleaned by a guy in a "hazmat" suit with a pressure washer. It probably wouldn't be that hard to largely automate this. But the market for this is small and mostly governments, so prices are high and innovation is low.
Nobody said no access to porn harms children. They said that mandatory age verification, primarily but not exclusively targeting porn, has side effects that harm children, which is honestly very predictable.
Go look at some of these laws around the world, or especially the proposals that didn't make it through the first round and their proponents are trying again. It's never just porn.
And what, even if they don't have documents through full fault of their own, they shouldn't be able to access information and means of communication?
And assuming everyone you care about is able to get documents, do you think they shouldn't be allowed to access information or means of communication anonymously?
The argument about people who can't prove their identity is a distraction. The argument should be why are we requiring anyone to do that in the first place?
To take it one step further: how does the election authority convince everyone that their vote was counted?
Us "computer folks" convincing ourselves is the easy part. Maybe after enough studying of the system, I'd be absolutely sure that my vote was counted, but what would the other 90 % of the population think? This not a "people are stupid" argument, a brain surgeon or nuclear physicist or philosopher isn't going to put in the effort to understand fully homomorphic cryptography, hardware-based security, etc. Why should they?
The alternative works perfectly fine and it's so easy to explain. It's obvious to anyone, no matter their level or field of education, how their vote makes it from their hand into the tally, what prevents it from being misused and what they can do to further reduce that already tiny probability.
You can see, feel and experience the technical measures keeping your vote safe. The people keeping their eyes on the ballot box alp day may literally be your neighbours, as are the people counting. You can reason about their motivations, you can theorise in your head about what lengths you'd have to go to in order to change even a few votes and how unlikely you would be to succeed.
The biggest danger to elections, ultimately, isn't the risk of cheating. It's the risk of losing the people's trust.
(I'm in the local election commission (that's layer 2 out of 4 in our particular election management hierarcy), we just had a big referendum, everyone keeps asking me about online elections so this rant was fresh on my mind)
I'm really not worried at all about if my vote is tallied correctly. I'm more worried about mass fraud of the voting system.
It's funny too - 15 years ago on HN people would discuss blockchain voting systems or ways to make it 100% secure. It seems like lately a lot of people carrying water for the potential that one party is associated with 95% of the fraud - and don't wish to discuss the vote being fair.
reply