0 sats \ 0 replies \ @note_bene 27 Mar \ parent \ on: Vatalik Buterin's MASSIVE $665 Million Donation to AI Nonprofit Makes Waves AI
Agree 100%. Why can't there by an e/acc charity that takes in a billion? What isn't there an e/acc channel on stacker news?!
A lot of people are catching on to this school of thought.
Some interesting ones of note:
- MacKenzie Bezos to Social Justice Inc
- SBF to Anthropic (which possibly led to getting enough compute to make Claude3 on par with GPT4)
- Jack Dorsey to various Anti-Racist / BLM initiatives
- Dustin Moskovitz to EA
- (what else?)
Beyond just today's headlines, there seems to be a long history of the non-profit sphere being infiltrated and subverted by a progressive agenda - Ford and Rockefeller foundations are usually the go-to examples. But overall, even twenty years ago it felt like charitable giving was more centrist where something like United Way - which would fund a mix of both religious-type and social worker type approaches - was much more common than today. But maybe that's because the super rich were more traditional industrialist types, rather than today's technologist / futurist set.
What are these mega donors buying for their billions? I would speculative psychological motives as: MacKenzie wants to be liked by cool people, Dorsey and Buterin want to be principled and visionary, SBF and Moskovitz want to expand their sphere of influence.
Beyond whatever policy recommendations are promoted by FLI, I think it's worth examining what makes these nonprofits so seductive to their benefactors mindset when choosing which of the thousands of organizations that could use the money.
Hmm, cashapp gets a different error when paying an lnbits or thunderhub generated invoice wrapped with lnproxy, error code: "Invoice missing valid amount...Please create or requests a new invoice with an amount of at least 1 satoshi". For reference, phoenix and strike do pay proxy wrapped invoice.
Agree, lightning looks like it could have the "Mastadon problem" where it is decentralized (like Mastodon the app) but that just makes interop across all the users more difficult because every federation is putting different outbound/inbound restrictions on nodes they deem baddd, m'kay?
Yes, I have phoenix and use it as my primary.
One reason I sometimes use CashApp is sometimes phoenix gets stuck "Connecting to Electrum" and can't make payments for several minutes.
But the larger point is I'd love regular old folks to pay me via CashApp instead of instructing them to take an extra step with a secondary wallet. Beyond the friction of extra steps, if the user wants only like $5 worth of services, Phoenix is also going to take a good chuck of say that for channel opening fee, which will seem predatory to a first time lightning user.
For a.) reliability and b.) QA 'ing the potential to onboard non-lightning people to my app, see reply below to kepford.
Funny enough I have that tab open, and just reading about when I came back here to check for replies. Thanks for the advice! I'll report back here with what I find.
Well said, i think you'd like sms4sats, one purpose, useful, good ux.
Most of all, something you can only use sats for, which is the main thing lightning needs to offer to gain use beyond early adopters.
And in the background, that's something that nostr could be setting the ecosystem up for by allowing arbitrary peer to peer routing and being the main way to route around censorship, centralization, and surveillance.
Sms4sats is solving this problem in miniature, a way around sms verification, except nostr is creating a whole alternative system that doesn't need sms verification measures, but as you've said this is somewhere well down the road.
Interesting,...And yes the inbound capacity seems fine, especially for these small 10-50 sat payments. I can pay other wallets with cashapp, and I can pay my node with other wallets.
I was going to ask about workarounds, and sounds like BTCPay Server is a good one. I was also wondering about lightning addresses, and whether you could create arbitrarily many of them to always generate a new recipient for each invoice or something like that?
Yep, webassembly is giving web and js devs access to the power of the gpu, such as your example of 3D rendering in realtime. This creates the potential to run ai inference on the edge (on user's computer) with cross compatibility across platforms and browsers, meaning no need to target each platform - iOS, android, windows - with a native app.
The other potential that webassembly - which I know less about - is enabling safer and more secure application, maybe delivered as progressive webapps. If i recall correctly this is something supertestnet was talking about on a podcast - write a js implementation of lighning with webassembly - and something Mutiny wallet might be doing?
See this is why I told k00b to build on Solana.
Had he done that, the network would have been down and he never would have lost the coins.
On a serious note, great postmortem. Curious for the Typescript folks, does TS actually protect against this? If so, how?
Finishing up a project for the Boltfun hackathon. Checking out some of the projects.
Planning out the next month for personal and dev growth.
UTXO makes great point about upside/downside tradeoff for merchants:
The indifferent to mildly curious (majority of people) don't want cryptographic sovereignty and see thus see little payoff to the friction that creates in the maintenance and setup. Instead, they just want a frictionless check out option, and to be paid reliably and conveniently.
Nodeless seems to be addressing the right concern, for a large potential mass of users; nice work.
This whitepaper is going down as an all time classic. But then again I'm biased because I've been thinking about it for awhile...
First, LightningAPI's for AI services is the definitive answer to the objection "What can you even buy with bitcoin?" Similar to SilkRoad days, the main thing people will want sats for, is things you can only buy with sats. And since OpenAI, Microsoft et al. will be pressured by their compliance departments, the most powerful and most creative uses will be open source, but difficult to run for the average users (software skills, access to hardware), thus lightning-AI.
A couple of overlooked ideas in the thesis for serving models:
- EC2 for GPUs - Generative AI artists have highly variable demand for image generation: sometimes they want access to multiple GPU's at once to generate a bunch of different draft images or render frames in a video, and sometimes they are doing just touch-up and paying for full use of even 1 GPU is overkill. Existing services (or self-service) will not allow you to both pay-per-use and scale up to multiple processing units on-demand.
- This brings up another intriguing idea...the AirBNB GPU strategy, buy to fulfill your own need but also to rent spare capacity. This strategy was mentioned yesterday by GeoHot for his new AI server which have "FLOPcoin" apparently .
- Etsy for Models - Currently there's not a great way to monetize a fine tuned model, if you check out civit.ai people are uploading their stable diffusion models simply for clout. With a strong financial incentive, this could supercharge this space, and give the most succeful fine tuners ways to invest into more serious training runs.
Maybe most exciting for the average StackerNews reader is potential for the much awaited bitcoin circular economy: Artists use sats to pay for gpu cycles, gpu suppliers use these sats to buy more servers/models/gpus, and Artists use the images they create to collect more sats from fans/customers. None of these participants even have to like bitcoin, just recognize it's the mana that ties this decentralized ecosystem together.
No - the CIA did not find Satoshi and throw him in Gitmo and waterboard him for his keys.
Yes - Satoshi believed that continued use of his established digital identity (forum account, email server, etc) with the CIA/NSA actively looking into him would result in his de-anonymization. This was two years prior to release of the Snowden documents, which shows Satoshi was savvy enough to know his both his own limits of op-sec and the extent of logging infrastructure that had been put in-place to ensnare people like him.
I guess if anyone could deanonymize Satoshi, it would involve subpoena or hacking forums on which he talked and getting the ip address, and then going to the ISP and again subpoena for who paid for that address. How long those logs are preserved and whether that leads to a real identity are unclear.