The new mayor of NYC wants the city to take over grocery stores. Very bullish for mechanism designers. Personally I'm excited to apply my experience working on Solana queues to NYC bread lines.
Experienced teams from across the industry who want to build efficient markets are reaching out asking about building with ACE on Solana.
Without revealing too much about our immediate plans (announcement coming soon) I can say that Solana development philosophy is to talk with apps, figure out what is going wrong, and fix the problems quickly.
If changes need to be made to make sure that Solana can be the home of the worlds most liquid markets, we will make them. Right now it looks like the market structure would actually work really well (at least considerably better than today) if we could just get good transaction landing. This is because of prioritization based on prio_fee/requested CU in the scheduler. To that end we have deployed the artillery (@alessandrod) to dramatically reduce landing latency and jitter. I'll report back with before/after metrics once the changes we identified so far land.
Experienced teams from across the industry who want to build efficient markets are reaching out asking about building with ACE on Solana.
Without revealing too much about our immediate plans (announcement coming soon) I can say that Solana development philosophy is to talk with apps, figure out what is going wrong, and fix the problems quickly.
If changes need to be made to make sure that Solana can be the home of the worlds most liquid markets, we will make them. Right now it looks like the market structure would actually work really well (at least considerably better than today) if we could just get good transaction landing. This is because of prioritization based on prio_fee/requested CU in the scheduler. To that end we have deployed the artillary (@alessandrod) to dramatically reduce landing latency and jitter. I'll report back with before/after metrics once the changes we identified so far land.
Fartcoin quants have been doing TAM calculations all wrong. They considered the total addressable market for digital natural gas based on the total quantity of human produced methane. but the largest producers of methane aren't humans they are livestock.
Open AI, Anthropic, Gemini, and Deep Seek are the frontier labs of AI. Which places in crypto are frontier labs today pushing the boundaries of the technology?
Anza/FD are obvious. Mysten certainly deserves consideration. Which others should be considered frontier labs?
Market structure influencers on twitter would have you believe that Solana cannot compete with tradfi because it is a decentralized network. The numbers don’t the support narrative.
Today Solana spot spreads are sub 1 bps, in many cases even tighter than Binance.
This is because of permissionless innovation from Solfi, Obric, ZeroFi and others. We are going to accelerate that by giving app developers even more tools to innovate on the market structure.
I'm running some transaction landing experiments rn and i'm sending about 20 tps. Which I realized is more TPS that the entire Ethereum blockchain. I started with 0.05 Sol in my account, now i have 0.04984 Sol.
Transaktionen sollten bei der Einreichung vollständig statisch sein.
Warum: Sequencer müssen in der Lage sein, über Transaktionen nachzudenken, ohne sie auszuführen. Und Validatoren sollten in der Lage sein, über die Gültigkeit von Blöcken nachzudenken, ohne den Block auszuführen.
Alle sequenzierungsrelevanten Eingaben, einschließlich, aber nicht beschränkt auf:
1. Gebührenzahler 2. Ressourcennutzung und erhobene Gebühren 3. Anweisungen 4. Lese- und Schreibzugriffslisten
Sollten innerhalb der Transaktion enthalten sein.
Tatsächlich sollte sogar, welche Konten geladen werden, statisch und nicht zur Laufzeit manipulierbar sein, um Cache-Misses zu vermeiden (Sie sollten sicherlich in der Lage sein, etwas, nach dem Sie gefragt haben, nicht zu lesen, aber Sie sollten nicht in der Lage sein, etwas zu lesen, nach dem Sie nicht gefragt haben).
Wenn der Sequencer die Kontosperren und Ressourcennutzung Ihrer Transaktion nicht simulieren kann, ohne Zugriff auf die Laufzeit zu haben, ist es zu dynamisch.
"Warum ist die Wartezeit für p90 bei Ping immer noch so lang?"
Ich denke, die meisten Menschen (die an Gebührenmärkten arbeiten) verstehen, dass die Wartezeit in einer FIFO m/m/1-Warteschlange mit Überkapazität langgestreckt ist. Das bedeutet, dass Sie die meiste Zeit ankommen, wenn die Warteschlange leer ist, aber ein kleiner Prozentsatz der Zeit ist die Warteschlange voll, und wenn sie voll ist, ist die Wartezeit in der Erwartung super lang, wenn dies geschieht.
Was die Leute vielleicht nicht erkennen, ist, dass dies auch für Prioritätswarteschlangen für jede gegebene Prioritätsstufe der Fall ist. Angenommen, Sie zahlen eine Prioritätsgebühr, die höher ist als 30 % des anderen eingehenden Verkehrs. Dann ist Ihre Wartezeit in der Prioritätswarteschlange mindestens so lang wie in der FIFO-Warteschlange mit 70 % Auslastung und wahrscheinlich höher und viel langgestreckter, da, wenn Sie in der Warteschlange sitzen, alle Transaktionen mit einer höheren Prioritätsgebühr, die ankommen, vor Ihnen gehen.
Solana is not focused on CLOBs, FCFS, cancels before takes, or any other market structure. There are no Zealots here, only pragmatic people who want the tightest most liquid markets to exist on Solana. Whichever market structure leads to that is fine with us. But the reality today is that if an app wants to expirement with some new market structure Solana isn't the place to do that. Applications don't control their own sequencing on Solana, Validators do. That is not going to work long term.