Jesus and Bitcoin. Building an operating system.
Joined March 2007
- Tweets 11,505
- Following 1,656
- Followers 25,168
- Likes 11,814
621 Photos and videos
Jun 12
I like Fable. Anthropic's reluctance to share and un-neuter this model generation will be a huge win for open source. And for a couple weeks I get a nice preview of what ultrabig models will excel and fail at it. I'm having a good time. I refuse to cope and seethe.
1
1
4
742
Paul retweeted
Jun 3
Wow
I’m so humbled to see Finite mentioned on the main stage at @OsloFF by @gladstein and Noemi from The Reynolds Foundation.
This is the coolest thing I’ve ever worked on and I can't wait to reveal our vision for AI. Stay tuned...
Watch this clip 👇
4
6
22
3,483
Paul retweeted
May 19
New version of Nostr VPN is out. Like tailscale, but no email addresses or 3rd party accounts, just public keys. New:
* native multiplatform user interfaces
* Nostr-based multihop routing (FIPS protocol) — very useful when NAT holepunching fails
* improved network management UX
65
272
1,534
164,289
Paul retweeted
13 May 2020
In the pyrite community we refer to gold as "fool's pyrite"
40
5,147
38,967
Paul retweeted
May 13
The Human Rights Foundation would like to fund more self-sovereign AI projects.
Here is the link to apply and I can answer questions hrf.org/program/ai-for-indiv…
5
9
62
4,910
May 11
The fool mistakes power for control.
"Though kings seem to have all things at their command, yet they are not able to bring their own purposes to pass any other way than God has appointed. Much less are those below them able to do so." (Geneva Bible on Proverbs 21:1)
2
4
445
Paul retweeted
May 11
If you use a variety of open models long enough, you inevitably run into all sorts of problems.
Most issues are not model intelligence issues. they are silent errors like outputting no content, malformed tools, etc.
Today I'm releasing Attune, a new PoC to try to solve this.
2
4
15
1,380
Paul retweeted
May 6
You know that hesitation before you type something into AI that you might regret? We built Maple to make it disappear.
Today we're (re)introducing Maple, the Personal Intelligence Platform. Encrypted AI for your real life.
5
34
124
31,711
Apr 2
Tonight!
The train keeps on rolling!
Join us on 4/2 at @bitcoinpark_ Austin for our ABC continuing to build on our conversation of how AI is advancing innovations in the Bitcoin and Freedom Tech ecosystem.
We are pumped to be joined by @futurepaul, @nk1tz, and @SahilC0.
RSVP below.
1
5
923
Paul retweeted
Mar 20
I've been busy lately, just found some time to work on my side project again
So far I'm almost finished with this @fedimint overview in @ratatui_rs
I'm experimenting with using syntax highlighting for certain terms to make it easier to understand
2
7
48
4,109
Paul retweeted
Mar 19
Announcing the world's largest paid endpoint directory for AI agents: the 402 Index!
The 402 Index aggregates 15,000 paid API endpoints live on the internet today, and runs hourly health checks to help your agent determine which providers have the most reliable infrastructure.
73
78
496
77,266
Paul retweeted
Mar 18
1/
One line of code. That's all it takes to add a @mpp compatible lightning payment paywall to any API endpoint now.
No accounts. No API keys. No third parties. Just bitcoin.
Here's how I built it 👇
5
26
133
10,857
Paul retweeted
Mar 16
We now have a mutinynet cli tool! This should make using the faucet much easier, great for agents too!
github.com/benthecarman/muti…
1
6
37
2,930
Point your favorite AI model at api.zaprite.com/llms.txt if you're looking to integrate bitcoin payments into your app. Less searching and fiddling. Stack more sats.
3
10
45
6,167
We believe that using Bitcoin as money must be at least as convenient and fun as using fiat.
So we built the best Bitcoin (or crypto) payment UX in the world. Something our grandma could use.
Faster than tap to pay.
Multi-currency.
Privacy-preserving.
Free and open-source.
Mar 5
One tap.
Paid. ✅
This is what paying at a Numo terminal with your iPhone looks like.
24
59
654
45,449
Paul retweeted
Mar 4
Agents are what computers were always meant to be
5
8
78
5,768