whenever someone in crypto hits me with a ‘people will never care about privacy’ I wonder what their understanding of the ethos behind decentralization is
look around, the world is changing and we need to avert dystopia
drop the doomerism and believe in something
someone should start a fund that sits in crypto conference bathrooms and only invests in dudes who wash their hands because bros wtf
I have to shake your hands you gremlins
tokens create ux complexity when chaining infrastructure together
for the
@nillionnetwork
integrations with partner chains like
@arbitrum
@Aptos
and
@NEARProtocol
we're working hard to try and give ecosystem devs and users an experience that's as native to their eco as possible
when explaining Nillion to someone for the first time the goal isn’t to convince them, it’s to educate them
blind computation shills itself if you can get them to understand it
work to get first timers from:
>‘I don’t get it’ / ‘but what’s it for?’
to
>‘woah that’s magic’
been really great to work with
@danilowhk2
!
dude is cracked, showed up to our first in person meeting in Singapore with printed out diagrams for us to jam on
Nillion 🤝
@0xZapLab
Check out their latest blog post to learn more about how Zap aims to create a new category of valuable data pools and how Nillion unlocks the data's full potential
Hawk Tuah recently went viral for her rant on Privacy Enhancing Technologies (PETs):
"They’re just too hard to use today, that’s really a bottleneck for adoption. Plus there’s no legibility for how the trust is distributed for most people and where the tradeoffs are."
this is a big opportunity for
@nillionnetwork
infrastructure, we already have a few teams working on different relevant angles
if you’re working on building with sovereign data hit us up and take advantage of those network effects!
Private Datasets will be the next leap in AI models.
Data ultimately determines the quality of a model. Not compute.
Generally accessible data will produce general LLMs. They’ll answer your question or prompt with a broad, unspecific response.
Unlocking AI’s true potential
Data-to-earn is a really exciting space that I’m glad is picking up steam because people should be in control of their digital worth rather than platforms
…as long as we always remember to respect user privacy with strong PETs
is it subway surfer brain rot that in airplanes I like to watch the 4+ seatback movies happening in my field of view at the same time?
the opposite of rawdogging the flight
ps: thank you to all of those who put on subtitles
An important lesson we've learned through constant failure and iteration at Nillion:
stop trying to only sell that new piece of math or algo or whatever
people want to know how it will help them ___.
build things, protect themselves, make money, live better, whatever.
what percent of commercially useful AI operations are entirely transparent in web2?
basically none.
people rely on trusted third party cloud providers to not take a peek - but otherwise people don’t want the whole world seeing their data, prompts, or often even models.
the vision is a world where the digital world has the same expectations of ownership, privacy, and individual autonomy that we know to be Good and Important in the physical world.
until we make it there, keep building.
the world is not ready for how easily AI will make data accessible
stuff that was already available on the internet with a little bit of elbow grease will be pulled to your retina with 0 thought
we need better infrastructure and privacy enhancing tech is how we get there
My wife just texted me a key to her car, which is I guess the sort of thing you can do when you have the infrastructure for end-to-end encrypted messaging in place.
more people in crypto should be aware of Tim Berners-Lee’s Solid project
the data sovereignty it’s designed around speaks to the core of what this space can represent beyond the financials
there are definitely nitpicks, but the biggest flaw imo is that his distaste for crypto
work on stuff that you genuinely believe a better world should have
too easy to be snarky or jaded in crypto these days, but it doesn’t have to be like that
“The tech giants were spying on us and created mass surveillance infrastructure unchecked? Wow I didn’t know that, you’re telling me now for the first time.”
FTC staff report finds large social media and video streaming companies have engaged in vast surveillance of users with lax privacy controls and inadequate safeguards for kids and teens 1/
Durov getting arrested is very bad
Telegram’s approach to ‘privacy’ is very bad
Your favorite country is just as anti-privacy as all the others and this pearl-clutching is misinformed
We need better infrastructure
I'm playing with the weirdest new viral social app. It's called SocialAI.
Imagine X/Twitter, but you have millions of followers. Except the catch is they are all AI. You have 0 human followers.
Here's how it works:
You post a status update. Could be anything.
"I'm thinking
To monitor your cardiovascular and respiratory health, you have OURA, WHOOP, Garmin, Apple Watch, and dozen other options.
To monitor your gut health and hydration, you have
@ThroneScience
I am so freaking proud of what we've built in just over a year.
Today, we are publishing the Collective Intelligence Project whitepaper, our R&D agenda for transformative tech governance.
It is our contribution to what we see as one of the world’s biggest problems. We couldn’t be more excited to share it with you. 1/
it’s so jarring to look to the left and right of me on the AI spectrum and see that they both sort of horseshoe into ‘AI is evil’ but for opposite reasons (it’s too dumb vs it’s too smart)
meanwhile the rest of us are over here fiddling with the magic talking sand boxes
talked to some interesting people building stuff at Token - if you are or are interested in building hidden information games, like card or social deduction games, hmu
kinda wild that the universe of onchain games has been building without private shared state
feels like the design space is waaaay bigger once you have that
a lot of people working on making LLMs sound more human
funnily enough the reverse, using the LLM to make you sound more generic, is potentially useful anonymity tooling
like a voice changer for your online presence
my laptop broke in a very strange way and they basically replaced every part except the screen when they just repaired it
there’s no way the economics of that make sense though right?
kinda wild that the universe of onchain games has been building without private shared state
feels like the design space is waaaay bigger once you have that
users probably aren't willing to deal with the inconvenience of every aspect of their lives being financialized
but that information would be useful, the same way all markets are (just look at
@Polymarket
)
an information economy built on personalized AI agents is the answer
PayPal is starting a new ad network by selling your transaction data to businesses so those businesses can target you more efficiently.
Why is consent assumed?
And if your attention is valuable because of your own purchase history, why should PayPal be the one that benefits?
I just hope we get another cycle where people are excitedly reading the docs and being optimistic about what this tech can enable
So much cynicism in the air
RIP Social Apps, 2005-2023
As of iOS 18, friend-based contact sync apps are basically dead on arrival. The number of people consenting to the new permission has nose-dived—so there is no way to get density in a meaningful way.
But as one door closes, another opens up:
From
when explaining Nillion sometimes people ask:
'but why would users care about security now when they've historically chosen convenience?'
the friction of security and privacy has generally been too high and the market bulldozed the need, except in notable cases like TLS.
the
Tobi Lutke explains what the VCs who passed on Shopify got wrong
Tobi recounts pitching Shopify to VCs on Sand Hill Road a few years after founding Shopify.
Investors passed because they thought the addressable market was too small. At the time, there were about 40,000-50,000
trusting a single entity gives you infinite capabilities (at high security cost), trusting nobody gives you perfect security (with aggressively limited capability)
there's a whole world in between the extremes where most actually useful application will live in the future
@TrustlessMike
@nillionnetwork
Blind computation is focused on a different part of the stack - data.
The notion of ‘privacy’ in crypto has been pigeonholed over time to mean transaction privacy, but the programmability of blind computation is intended to be much broader than that.
Fear is healthy because it what keeps us from creating unrecoverable scenarios
Opportunity is the much sweeter nectar, though. AI/ML can and will do amazing things if we get this right – supporting Nillion is recognizing that it’s all of our job to help us build that reality.
Many are suddenly afraid of AI for privacy but govts& companies have long used ML for surveillance. The change is that everyone now has access to AI, shifting control from powerful entities to individuals. It's a good thing, not something to fear just because you're aware of it.
the convergence is particularly weird because it doesn't feel like it's converged on things that feel particularly prevalent in the dataset – why are we delving everywhere?
Not fully sure why all the LLMs sound about the same - over-using lists, delving into “multifaceted” issues, over-offering to assist further, about same length responses, etc. Not something I had predicted at first because of many independent companies doing the finetuning.
“Any sufficiently advanced technology is indistinguishable from magic.” I wonder if for many people thinking about future tech as “what you could do with magic” rather than an extension of current tech would actually improve their model of the future
imagine training on a supercluster of everyone’s laptops and gaming pcs and homelabs
now imagine those devices training on data that even the frontier AI labs don’t have access to. the data we keep so close to our chests it isn’t scrapeable, made useful but kept private
What if you could use all the computing power in the world to train a shared, open source AI model?
Preliminary report:
Nous Research is proud to release a preliminary report on DisTrO (Distributed Training Over-the-Internet) a family of
glad that Chain Abstraction is gaining steam, it’s the only way to fix the UX snafu we’ve created. MPC is going to be a big part of that (and we can even sneak in some native privacy alongside the necessary stuff)
damn the old guard Berlin scene is getting rekt Watergate announced it’s closing. so many of the classic clubs closed in the last couple of years, sad to see
1- apps get users to store high value data with Nillion
2- users store different data across different apps
3- that data is composable and so more valuable, but also even more sensitive, stacked together.
4- blind computation lets everyone have their cake and eat it too
blind compute is pretty evocative of what it does but I’m still toying with a label for the applications that gets people to instinctively understand what kinds of cool stuff can be built with PETs - any ideas?
I don’t think ‘blind apps’ gets there
people I talk to about MPC forget that the CURRENT STATUS QUO is to trust the faceless employees of a giant corpo with your data
some use cases will only want to split trust across 2 node clusters, others 5, others 100 - ALL of which would be way better than where we are today