As others mention, show the keys in the interface. It took me a while to discover accidentally, by pressing space, that I can trim the sails. You could make a graphical guide, showing a ship with arrows around, some arrows attached to the hull with keys responsible for turning the ship, some arrows attached to sails with keys responsible for trimming them. Yet another key by the cannons fire to show which key... umm... fires the cannons...
I'm slowly trying to build an app on Iroh; it's progressing tiny bit by tiny bit, but I must admit I'm struggling a lot all the time, both with various low-level details, as well as with understanding many high-level aspects, concepts, and approaches. Oftentimes I have to resort to some LLM-generated "wiki" websites to help me progress. I really hope you'll manage one day to allocate some more resources to improve the docs. That said, when I manage to muster enough strength, I do manage to grind some progress, and also it's good to know the underlying tech seems robust, given how many real-world solutions you've built on it!
At this moment, if I can try to ask one question: AFAIU Iroh emerged from an attempt at fixing IPFS. I also understand you've since focused more on providing the lower-level building blocks that would allow this and other solutions. Understanding some basics, but still having hard time to get a really solid grasp of the whole of Iroh, I wonder: between Iroh, p2panda, and Willow, what's available and what's missing / needs to be added if one wanted to try and build an "IPFS-like" with those technologies? I'm especially interested in an idea of a "new web" that would defuse DDoS of static websites in a Torrent-like way, forcing the downloading peers to also share their upstream bandwidth while doing this. I'm also thinking of e.g. a "globally-distributed Internet Archive", where I can easily download part of the Archive to my computer, and this automatically improves its availability on such "new web" for subsequent downloaders and browsers. Would you care to give a newbie something of a high-level overview of how one could try to do it over maybe some appropriate combination of Iroh+p2panda+Willow+DHT?
We started as an IPFS implementation, but since then the scope of iroh has been reduced. Iroh is not IPFS, but more like libp2p.
If you want something like a globally distributed internet archive you would have to use protocols on top of iroh. For example iroh-blobs provides verified streaming of content-addressed data using the BLAKE3 tree hash function. It is very close to itself being 1.0 (probably Q3), but for now it requires you to know from where to stream the data.
What is missing to fully replace IPFS is a global distributed content discovery system. This is a really hard problem that IPFS itself never solved reliably in my opinion.
I also still want this to happen eventually, but the first step is to get the connection layer super reliable and fast, which we have done.
I wish you could also delegate this problem to the mainline DHT, but alas that is not possible because of some mainline limitiations. So I am working on the side on a new DHT, see https://www.iroh.computer/blog/lets-write-a-dht-1
Is content discovery on the official roadmap? If the original peer say peer A goes offline but B downloaded $X, can see find B and download, validate the content today?
Sorry if this has been answered or covered on the web I'm currently travelling.
Edit: I managed to read the linked page, thank you for working on such an amazing project!
We don't have the post 1.0 roadmap very fleshed out at this point. The team has been extremely focused on getting 1.0 out of the door for the last year. This took longer than expected, as expected :-)
But I can say that some form of global content discovery is in my personal definition of done for the project. I'm the BLAKE3 guy at n0, and I don't consider blobs done without global content discovery.
But it is a hard problem, and our standards for "it just works" are pretty high.
Excellent I look forward to it, I like Iroh objective of keeping it's footprint small enough to run on low powered devices. IPFS in my experience is very heavy.
If my understanding is right, content discovery can't happen on the existing mainline DHT which is a shame. Maybe once you have enough traction you can propose a standards change? DHT mainline today is amazing because of its network size.
We would love this to gain more traction so we can use mainline also for content discovery. But we have to be realistic: mainline moves slowly, and for good reasons.
"due to their inherent stochastic nature, there would still be a small likelihood of producing output that contains errors"
This is the part that I find challenging when trying to help my friends build a correct intuition. Notably, the probabilistic behavior here is counter-intuitive: based on human experience, if you meet a random person, they may indeed tell you bullshit; but once you successfully fact-checked them a few times, you can start trusting they'll generally keep being trustworthy. It's not so with "AIs", and I find it challenging to give them a real-world example of a situation that would be a better analogy for "AI" problems.
In my family, what worked (due to their personal experiences), was an example of asking a tourist guide: that even if the guide doesn't know an answer, there's a high chance they'll invent something on the spot, and it'll be very plausible and convincing, and they'll never know. I'm not sure if that example would work for other listeners, though.
I also tried to ask them to imagine that they're asking each subsequent question not to the same person as before, but every time to a new random person taken from the street / a church / a queue in a shop / whatever crowded place. I thought this is a really cool and technically accurate example, but sadly it seemed to get blank stares from them. (Hm, now I think I could have tried asking why.)
Yet another example I tried, was to imagine a country where it's dishonorable, when asked about directions in a city, to say that you don't know how to get somewhere. (I remember we read and shared a laugh at such an anecdote in some book in the past.) Thus, again, you'll always get an answer, and it'll sound convincing, even if the answerer doesn't know. But again, this one didn't seem to work as good as the travel guide one; but for now I'm still keeping it to try with others in the future if needed.
PS. Ah, ok, yet another I tried was to ask them to think of the "game" of "russian roulette". You roll the barrel, you press the trigger, nothing happens. After a few lucky tries, you may get a dangerous, false feeling of safety. But then suddenly you will eventually get the full chamber.
I also tried to describe "AIs" (i.e. LLMs) as taking a shelf of books, passing them through a blender, then putting the shreds in some random order. The result may sound plausible, and even scientific (e.g. if you got medical books, or physics textbooks). The less you know the domain the books were about, the more convincing it may sound, and the harder it is to catch bullshit.
The last two pictures may have gotten some reception, but I'm not super sure, and there was still arguing especially around the books; and again, they were less of a hit than the tourist guide story.
I'm super curious if you have some analogies of your own that you're trying to use with friends and family? I'd love to steal some and see if they might work with my friends!
AFAIU, their claim is that Mythos is in reality used in a framework that builds such contextual hints, and that their (Aisle's) own framework does the same:
"(...) a well-designed scaffold naturally produces this kind of scoped context through its targeting and iterative prompting stages, which is exactly what both AISLE's and Anthropic's systems do."
All evidence is point to LLMs not being sufficient for the taks everyone want them to do. That harness and agentic capabilities that shove them through JSON-shaped holes are utterly necessary and along with all the security, that there's no great singularity happening here.
The current tech is a sigmoid and even using the abilities of the AI, novelty, improvements don't appear to be happening at any exponetial pace.
