1
point
rglullis
Joined 12,255 karma
Some of my projects:
- Communick (https://communick.com) is a provider for social media and messaging platforms that are based on open protocols.
- CareerCupid (https://cupid.careers) is a site to make it easier to see if job seekers fit well with a company and would-be coworkers. Like a dating site, but for your professional connections.
- Django ActivityPub Toolkit (https://activitypub.mushroomlabs.com) is a Generic ActivityPub server that can connect you or your existing web application with the wider Fediverse.
- Fediverser (https://fediverser.network) aims to make it as easy as possible to migrate from Reddit to Lemmy.
- Hub20 (https://hub20.io) is an open source, self-hosted payment gateway for cryptocurrencies, currently focused on Ethereum and its scaling solutions. You can accept payments and also use your node to host "crypto wallets" for your friends or a closed community of people that trust each other.
[ my public key: https://keybase.io/lullis; my proof: https://keybase.io/lullis/sigs/BInOa9RtBJ7qEX9m-kep0dQDBGxoNb5rXzhno3M27ok ]
The best way to reach out nowadays is through Mastodon (@raphael@communick.com) or Matrix (same username/server).
meet.hn/city/de-Berlin
Socials:
- github.com/lullis
---
- rglullis parentOooh... can you imagine if servers actually took the hint and sent only text if the client provided Accept: text/markdown, text/plain headers?
- Then you'd have to explain why every reputable newspaper is putting up paywalls and all the quality sites that used to cover specific niches went out of business in the last 20 years.
Saying "revenue goes up with content quality" only makes sense if you compare one poor site with another, but when you put in terms of ROI, you will see it is a lot easier to set up hundreds of different content farms than to keep a sustainable source of well written, reputably sourced reports.
- > A funding crunch since 2023 yet those features have been necessary for many years before 2023.
But before 2023, the funding was going to things like solving state resolution, a VoIP system that was not dependent on Jitsi, getting rid of "could not decrypt message" errors, and so on.
- > I dont see any reason why an app approach cant support that.
Matrix clients have e2ee encryption like Signal or WhatsApp.
Every single one of my close contacts that I have on my server have ignored or misunderstood the instructions to download and store the recovery key when they first access the servers.
I have customers on my support channel who keep trying different clients (Element, ElementX, Fractal) and every time they fail to validate their sessions.
Then I have customers who got their phone stolen and then come asking me to either delete the data on their phone.
---
There is no magic about "putting it in a app to manage it". If any "app approach" you come up with creates a sandbox between user and device, then the user can not even see their private keys, then they effectively do not own it.
If you are doing "nostr, but with keys sandboxed on the device", then you are just recreating Signal - which is not decentralized - then what's the point?
- > No just the first part,
Ok. Then the question is "If the US really goes full isolationsist and packs it home, who takes the power vaccum?"
> China ain't doing no ground invasion for oil
I didn't say that "China handling it" is about invading anything. I also didn't say anything that the US is justified in invading Venezuela. I am just wondering that if those saying "the US shouldn't do anything" understand that someone will do something, even if this something is stupid, counterproductive or plain evil.
- My point is that is this is not a trade-off but a complete violation of the principles that are used to justify the existence of nostr.
Nostr's whole shtick is about "users owning their keys". If I can not change the keys used on WhatsApp or Signal, I do not own them. They are not in the same class, so the comparison is moot.
- It is not about the difficulty, it's the potential consequences.
People also take care of their house keys and their wallets, but If I lose the keys to my house, it isn't automatically taken over by squatters and if I lose my ID card I can issue a new one quickly.
What happens if you lose the cryptographic key to your nostr account? Who do you call for help?
- That also goes to the other extreme.
For all the faults of current Fediverse software implementations, it at least gives more options than nostr. If you don't care about controlling your own identity, you can use someone else's server. Nostr doesn't give you that, it's all or nothing.
- I'm again toying around with the idea of building an ActivityPub Server built around the principles of RDF, JSON-LD and the Linked Data Platform. [0]
It can work already as a "Generic" ActivityPub server and it can be made to work with Client-to-Server API, but given that there are not mature clients for that, I am now in the middle of an exercise where I am taking the existing server and implementing Lemmy's and Mastodon's APIs based on top of it. Once I can get any Lemmy and a Mastodon client working, I will then start changing their own SDKs, and then I can replace calls from their application-specific APIs with direct calls to Linked Data server.
[0] https://activitypub.mushroomlabs.com - I don't know how else to get my point across: what I am trying to say is that there is nothing "smart" about an automaton that needs to resort to A* algorithm implementations to "solve" a problem that any 4-year old child can solve just by looking at it.
Where you are seeing "intelligence" and "an existential crisis", I see "a huge pattern-matching system with an ever increasing vocabulary".
LLM's are useful. They will certainly cause a lot of disruption of automation on all types of white-collar work. They will definitely lead to all sorts of economic and social disruptions (good and bad). I'm definitely not ignoring them as just another fad... but none of that depends on LLMs being "intelligent" in any way.
- > Im saying the model is "intelligent enough" to solve a maze.
And I don't agree. I think that at best the model is "intelligent enough to use a tool that can solve mazes" (which is an entirely different thing) and at worst it is no different than a circus horse that "can do math". Being able to repeat more tricks and being able to select which trick to execute based on the expected reward is not a measure of intelligence.
- > when you can just plug their OpenAI-compatible API URL into existing tools?
Only the self-hosting diehards will bother with that. Those that want to compete with Claude Code, Gemini CLI, Codex et caterva will have to provide the whole package and do it a price point that is competitive even with low volumes - which is hard to do because the big LLM providers are all subsidizing their offerings.
- It's not the open models that suck, it's the infrastructure around them. None of current "open weights providers" have:
These are solvable without having to pay anything to OpenAI/Anthropic/Google.- good tools for agentic workflows - no tools for context management - infrastructure for input token caching - > They're not paying me to use it.
Of course they are.
> As long as the inference is not done at a loss.
If making money on inference alone was possible, there would be a dozen different smaller providers who'd be taking the open weights models and offering that as service. But it seems that every provider is anchored at $20/month, so you can bet that none of them can go any lower.
- Yeah, I am not interested in playing this tape again. The actions from Brendan as an individual are completely separate from his actions at Mozilla. Mozilla did not change any policy during his tenure and Brave is not accused of any discrimination practices or hostile to any minority group.
- My moral compass does not change based on who is being accused, but context is fundamental to make a proper judgment.
It is hard to come up with a situation where Google would be doing these types of tricks, because Google is already the dominant player in the market and they don't want to create products that cannibalize their own revenue streams.