paco bytes retweeted
Today we launched ihavenofriends.ai, a feed of characters you can talk to in realtime.
They respond in 500ms end-to-end (STT -> LLM -> TTS) on a single EC2 instance + Cloudflare Workers.
This costs ~$0.002 per minute of speech (90x cheaper than OpenAI's Realtime API)
paco bytes retweeted
Sorry, but this information is simply incorrect by orders of magnitude. (Just woke up, time to put out some facts).
SP1 is blazing fast and uniquely supports precompiles for keccak256 and full secp256k1 acceleration, giving it a 10x speedup that makes the costs of full ZK proving affordable.
OP Succinct's fixed proving costs are between $1000-$3000 a month, NOT $1M a year.
The proving cost scalability is also wildly off. Based on recent SP1 performance:
• ETH transfer: $0.0004 (4/100 of a cent!)
• ERC-20 transfer: $0.0007 (7/100 of a cent!)
• DEX swap: $0.004 (4/10 of a cent!)
This comes out to $1000s per month, NOT millions per year.
Not sure where the numbers in this chart come from (maybe its from benchmarking their own zkVM), but remember to DYOR and talk to us directly about OP Succinct and what it costs.
We have several teams who are already running OP Succinct testnets and have their mainnets scheduled to go live soon. The mission to make every rollup a ZK rollup continues.
paco bytes retweeted
Replying to @intoverflow
Tim, I really respect you and your engineering team. You all are really smart, which is why it's hard to feel like any of this is in good faith:
• A few weeks ago, we put out new numbers for proving costs. Members of your team replied to that tweet, so I know you guys saw it, and you could have cited those. xitter.cf/JacobEv3rly/status/184…
We have since improved SP1's perf further and gotten to the numbers I reference in my most recent tweet (release coming soon)!
• As @fede_intern mentioned, a blog post from 2 months ago in a space that moves very quickly is outdated (your team knows this very well):
xitter.cf/fede_intern/status/185…
He ran the stack himself and arrived at numbers an order of magnitude better than the above:
xitter.cf/fede_intern/status/185…
• Even when taking numbers from an outdated blog post, in the same table you cite, for Base we have numbers of 0.61 cents/tx. Why take the highest number out of all of them?
• Our post doesn't reference numbers for "fixed proving cost" (it's simplified, for readability), so I'm not sure where that number came from at all. Unfortunately, it's totally off.
In my reply, I simply wanted to clarify what costs we offer today, for any potential customers that might have the wrong info.
Taking a step back, all of this Twitter back and forth is just a waste of time for both of our teams and the space. We're focused on actually deploying our stack and having people use it. The costs I cited in my reply are what we offer to real customers today. And they're rapidly getting better.
Now that our info is out there, will try to not engage any further on this topic and focus on shipping, thanks.
paco bytes retweeted
(̿▀̿ ̿Ĺ̯̿̿▀̿ ̿)̄ ミ★ "𝘍𝘖𝘙 𝘐𝘔𝘔𝘌𝘋𝘐𝘈𝘛𝘌 𝘙𝘌𝘓𝘌𝘈𝘚𝘌" ★彡
TLDR: anybody.gg on @base NOW
𝘼𝙉𝙔𝘽𝙊𝘿𝙔 𝙋𝙍𝙊𝘽𝙇𝙀𝙈 is the most a͢d͢v͢a͢n͢c͢e͢d͢ t͢e͢c͢h͢n͢o͢l͢o͢g͢y͢ in the history of onchain gaming and ¢yptography.
paco bytes retweeted
1/ @topology_gg is hosting Autonomous World at @ethcc meet-up on July 11th, from 9:30 AM to 1 PM. Join to learn about the latest research findings in the Autonomous World community, including the latest updates on the Topology Protocol and other AW games. Link next thread.🧵 👇
paco bytes retweeted
I had the pleasure to keep contributing to an awesome project for the @Optimism Stack called Concrete.
Project: github.com/concrete-eth/conc…
PR: github.com/concrete-eth/conc…
Thanks for the opportunity and guidance @therealbytes 🫡
@OnlyDust_com
@Dojo_Coding
paco bytes retweeted
Curious about onchain games? Join us at 2.30 for a talk of @therealbytes from Concrete Labs on making real-time onchain games 👾
paco bytes retweeted
for(float e,i,a,w,x,g,h;i++<90.;o+=.01-.02/exp(max(s,e)*3e3)/h){vec3 p=vec3((FC.xy-.5*r)/r.y*g+2.,g);p.zy*=rotate2D(.5);e=p.y;h=e+p.x*.3;p.z+=t;for(a=.6;a>.001;a*=.7)p.xz*=rotate2D(5.),x=(p.x+p.z)/a+t+t,e-=w=exp(sin(x)-3.)*a,h+=abs(dot(sin(p.xz/a*.3)*a,r/r));g+=e=min(e,h*.5-1.);}
paco bytes retweeted
in order to understand the unlocks of onchain gaming, we must understand games from first principles
archetype.mirror.xyz/unOEXm_…
paco bytes retweeted
RetroPGF 3 is a wrap!
It's true some aspects of it could have been better, but the only people who never make mistakes are those who never do anything.
Huge kudos to the folks at @optimismFND for putting this together. It will have a big impact, and I have no doubt it will get better.
The Optimism grant program started chaotic, but it's now grown in a highly effective (if somewhat complex) system. It's by no means perfect, but it keeps improving, driven by the effort of the foundation & the initiatives of some truly heroic delegate (too many to name, but the people on the councils have been doing a great job for instance). We're now seeing RetroPGF follow a similar arc.
And importantly, this iterative improvement process was *always* the plan. We'll try something, and then we'll try something else.
So many thanks to the people at the foundation, many thanks to the citizens who took their task seriously, and to those that have promoted this endeavour or offered constructive criticism.
wagmi ☀️
paco bytes retweeted
I didn't even know there was a quorum - you might not know either.
Test in Prod is not passing the quorum for rPGF😢
As a core dev building OP Stack days & nights, it's a bit discouraging. Come on.
xitter.cf/elblockchainguy/status…
paco bytes retweeted
Test in Prod is not passing the quorum for rPGF😢
As a core dev building OP Stack days & nights, it's a bit discouraging. Come on.
xitter.cf/elblockchainguy/status…
Test in Prod (@testinprod_io) literally built & now maintain OP-Erigon, are an OP Stack core developer, but are only on 8 ballots right now.
IMO they are one of the clearest cases for RetroPGF.
vote.optimism.io/retropgf/3/…
paco bytes retweeted
How do you find a balance between designing a digital physics that allows for complex and unexpected behaviors to emerge and ensuring that the available infrastructure can support these?
The essay “Architecting Physics” by @therealbytes is now online:
aw.network/posts/architectin…
paco bytes retweeted
🔥 What an awesome journey we have with @optimismFND since the launch in September!
First, we are deeply honoured to collaborate with them after being selected with the best note amongst +30 projects 🤓 via their grant program 🥰.
In turn, we created a committee to distribute tokens to projects who were building value in the ecosystem!
Big shoutout to the 3 jury members that we chose - @norswap, @jack_anorak & @dcbuilder who were super reactive and supportive 🙌!
Thankful for their help in going through a number of projects to vote, which in turn allowed us to onboard 4 amazing projects onto Only Dust🔥❤️!
25K OPs were thus distributed to the following cool projects:
0x Fable — a fully on-chain trading card game, in the spirit of Magic the Gathering and Hearthstone.
Cannon RS — a tooling to develop and emulate verifyable programs compatible with Optimism Cannon and the OP Stack.
Concrete — a framework for building application-specific rollups on the OP Stack.
OP Network Analytics — An Open-sourced analytical dashboarding tool for usage and health of the network.
They are all looking for contributors so if interested you know what to do 😉 - app.onlydust.xyz/
And finally, when participating at @EFDevconnect in Istanbul, we were able to present all this to over 250+ devs from the world of Starknet and Optimism!
Stay tuned as other announcements are to follow up in the following days with new builders and ecosystems arriving in Only Dust.
There’s no stopping us now! 🚀
paco bytes retweeted
RetroPGF badgeholders: I've created a small list for projects that contribute to/enhance the OP stack:
round3.optimism.io/lists/0x4…
paco bytes retweeted
Mad props to @therealbytes for a great talk on AW performance but MORE IMPORTANTLY for discussing the difference between “inchain” and “onchain” and not hyphenating EITHER of them 😁






























