
I’m all in on Solana
✓Founder @gotokamai | retired @solana validator | @RadiantsDAO | pfp by @scumsol magi collection

I’m all in on Solana

Looks like a good spot to buy $NVDA What’s the best Solana app to buy stocks today?


6 days ago I opened a SIMD to halve the @solana VAT. Since then: → toly on the discussion: "I am in full support." → operators are posting their real numbers One of them runs 133k SOL of stake. 69k of it is Foundation stake. 5% commission, $1,500 a month in servers. Today he makes about 20 SOL a month. When the Foundation stake goes, he's left with 64k: ↳ VAT at 0.8: about −12 SOL a month ↳ VAT at 0.4: about break-even That's the difference between shutting down and staying online, for 1 validator that has run for years. If you run one, drop your numbers in the discussion. Every real P&L makes this harder to wave away.

The Space: x.com/i/spaces/1eaKbrykWnvKX The Wired piece: wired.com/story/ftx-serum-cr…

November 2022. I'm flying home from Breakpoint in Lisbon. I land, open my phone, and FTX is collapsing. And the upgrade key to Serum, the order book half of Solana DeFi ran on, is sitting with FTX. Why that was terrifying: whoever held that key could push new code to Serum. Every protocol built on top was exposed overnight. And with FTX getting drained, we had no idea who else might have that key. I was deep in Serum at the time, writing code for SolApe, a DEX built on top of it. So the code I was shipping every day was suddenly sitting on a time bomb. What happened next is still the best thing I've seen in crypto. Within days, devs forked Serum into a community-run version: OpenBook. → a multisig took the upgrade authority, so no single party held the key → the code went out fast so people could keep trading → the Solana Foundation stepped in to speed everything up My part was mostly coordination: ↳ helping teams move from Serum to OpenBook ↳ reading the code ↳ running Twitter Spaces to explain what was going on, live, to anyone who'd listen By Nov 16, OpenBook had passed Serum in daily volume (~$3M). I called it "the point of no return" in one of those Spaces. Wired quoted that line, and listed SolApe among the apps that migrated. Fun detail: I've never been in Wired under my real name. The 1 time I made it, it was as Ansel. What I took from that week: a bunch of strangers across every timezone, with no boss and no contracts, rebuilt a core piece of Solana in days. That's still the Solana I'm fighting for. So here's my next fight: a staking pool that sends stake to the small validators the new SFDP rules are hitting hardest. You keep your SOL, the staking rewards turn into prizes, and the longer you stake, the better your odds. 100% onchain. 0% protocol fee at launch. More soon. Small validators who want in: DM me.


28.83% vs 33.34%. The share of stake that went dark on Aug 12, and the line where Solana stops finalizing. 1 routing error at 1 hosting provider in Miami. ~90 validators offline for 33 minutes. Blocks kept coming, but finality had 4.5 points of headroom. That provider's network held 27.34% of all staked SOL that morning. The small operators are leaving: ~2,500 validators in early 2023, 698 today. Stake is pooling into big shops. And the big shops just showed they share the same racks, the same routes, the same single point of failure. Only 3 of 74 tracked validators failed over automatically that day. Some of the biggest names sat offline for the full 33 minutes. Fewer, bigger, professional validators only works if they're actually more resilient. Aug 12 says the jury's still out. The cheapest fix I know: make it affordable for small operators in odd datacenters to stay online. That's what halving the VAT does. What would you make mandatory for a top-30 validator?

Crazy tool @SOLBigBrain Recovered a bunch of Candy Machines and a XL Serum deposit still handing there OG recognizes OG, respect!

~2,500 Solana validators in early 2023. 698 today. I'd like to keep as many of them running as possible. If you run one, send me your numbers (reply or DM): ↳ active stake, and how much of it is SFDP ↳ commission ↳ monthly server cost I'll send back your P&L: → once your SFDP stake ends under the new 12-month rule → with VAT at 0.8 vs 0.4 Free, anonymous unless you want credit. The ones you're ok sharing go into the VAT SIMD discussion as real operator data. That's how this gets passed.

A Solana validator operator has asked to cut the fee every validator pays to vote in half. @ansel_sol, who ran the Tokamai validator, opened discussion #730 on 10 November: the Validator Admission Ticket would drop from 0.8 to 0.4 SOL per epoch. Since Alpenglow, that ticket is flat and burned, "whether you hold 5k SOL or 5M." The timing is the Solana Foundation shrinking its delegation program. Ansel counts 337 validators in it when the new rules arrived, roughly half the network. @toly replied: "The vat number is the made up one. It's synthetic." If an implementation bug blocks lowering it, he said, fix the bug. Anza's @alessandrod put the cost of computing rewards each block at about 3ms today, and said it can be brought under 1ms. The next step in Solana's process would be a numbered SIMD pull request. For now ansel is asking other operators to post their own validator numbers in the discussion.

SIMD discussion is open here, come argue with me: github.com/solana-foundation…

I paid $500/month for 6 months to test Alpenglow. A few weeks after it shipped, the Foundation cut my SFDP stake. And I absolutely agree with the move. Directionally. Some context: I turned off @gotokamai validator 2 days ago. Yesterday I put my 2-year validator P&L out. Toly replied: lower the VAT by 50%. @TheWattenhofer (who co-wrote Alpenglow) said it's worth considering now that the delegation program is changing. So I wrote the SIMD. SFDP was a subsidy. When the new rules dropped on Nov 6, 337 validators were in it. Roughly half the network, kept alive by ~5% of staked SOL. Ending that is healthy. But once it's gone, 1 number decides who can afford to run a validator: the VAT. 0.8 SOL a day, flat, burned, whether you hold 5k SOL or 5M. Roger's framing: the VAT is there to protect against "too many" validators. Alpenglow already caps the active set at the 2,000 highest staked. We're at 698. That's 1,300 empty seats before "too many" is a problem. On my own books: ↳ old vote fees: 2 years net 329 SOL, year 2 at −0.7 ↳ VAT at 0.8: net 487, year 2 around +118 ↳ VAT at 0.4: net ~830, year 2 around +275 Cost to the network: ~102k SOL a year less burned. About 0.45% of annual issuance. That's the price of keeping the operators who test upgrades and spread stake away from the big hosting providers. The ask: VAT × 0.5 at every slot-time stage. Same burn, same cap, same stake ordering. 1 constant table behind a feature gate. Operators: reply with your numbers (stake, monthly costs, what share VAT eats). Or drop them straight in the SIMD discussion below. The more real P&Ls in it, the harder it is to ignore.

Stopping your solana validators cons: - fewer independent operators - feeling less connected to solana Stopping your Solana validators pros: - can switch off ALL DISCORD NOTIS 🤠 - no need to think about testnet anymore It might be a great time to look at smaller setups now

Response to the Foundation on SFDP changes We appreciate the Foundation responding publicly. Conversations at Breakpoint and Block Zero are valuable, but public discourse lets operators who cannot attend those events be heard too. And there is real agreement in the response, so let's start there. The response says Solana is much larger than when SFDP started, that the validator set has matured, and that the network faces challenges nobody conceived of at the beginning. We agree. We said the same thing in our letter. Solana is entering its next stage, and delegation should evolve with it. Where we part ways is what that evolution looks like. If the network has matured, the goalposts should move. Instead, SFDP is being redefined to complete its original goal again. That is not the next stage. It is the first stage on a loop, except the new rules boot the long-term contributors who are the output of the first stage and replace them with short-term 12 month tenants. The network trades operators who stayed for years for operators who are built to leave. It regressed, not progressed. Supporting new validators and supporting existing contributors are not mutually exclusive. A mature program does both. These rules prioritize only one group, and it is not the existing talent. First Principles The question is simple: Why does the new program value a brand new validator over one that has been operating reliably for three or more years? New validators will continue to receive stake. Validators who have run through every upgrade, every outage, and every cluster test since 2023 will lose it. We have read the announcement and the response several times and cannot construct the argument that makes the first group more valuable to network health than the second. If that argument exists, we would like to hear it. Alpenglow Community Cluster In follow-up comments, the Foundation said the Alpenglow community cluster participants provided crucial help, that they should be directly rewarded, and that these changes exist in part to make that possible. We welcome that. It is the first time the Foundation has said it. But it raises more questions than it answers. First, none of this was communicated with the initial announcement. The announcement said: you are out, the offboarding tranches start now. Only after independent validators pushed back publicly did the Foundation say contributors would be rewarded. And when asked how, the answer was opaque. Operators who ran the cluster at their own expense did not know until this week that they were about to lose their delegation, and still do not know whether they will get any of it back. Second, the Foundation says that with so much stake tied up in SFDP there was no way to reward the cluster participants. That is hard to accept. SFDP currently holds over 10 million SOL in residual. Rewarding a group of cluster participants materially would not have required dismantling the program as we know it. Third, the people the Foundation now says it wants to reward are the people being retired first. If the intent is to reward them, the sequence should be reward first, then restructure. Not retire, then promise. So if participating in the cluster is seen as valuable, will they be rewarded? Why does the retirement proceed before that is defined? "We reward contributors" is still not a program One of the responses says the intent is not to make running a validator a secondary activity, but to directly reward those improving network health. We agree with that framing entirely. It is close to word for word what our letter proposed. The response also says validators who contribute to network health should be rewarded, and points to past discretionary delegations to Blueshift and early Frankendancer operators as proof. We agree those were good decisions and believe discretionary delegations should continue to exist. But a handful of discretionary delegations is not a system. In addition, intent is not process. There are still no criteria, no application, and no timeline. Most contributors did not know discretionary delegation existed until this response. Now that they do, they have no idea how to access it or whether they qualify. "We staked to these guys before" is not an answer to "how do I survive on solana." Meanwhile contributors are shutting down nodes and cancelling Breakpoint trips because the only concrete thing they have been told is that their stake is leaving. This is not an argument against discretionary delegation. It is an argument for giving contributors something concrete to grasp to. Objective criteria show talented operators what the network values and can grow the contributor count. Discretionary delegation catches what the criteria miss. Both together work. Discretion alone asks everyone to trust a process they cannot see. On the feedback Asked to publish the feedback behind these changes, the Foundation pointed to a Discord channel where SFDP validators call out malicious or harmful operators. We know the channel. The people in it do real work. But that is not feedback on program design. It is validators reporting bad actors, and if the problem was bad actors, the obvious response was to bring those validators in and use their tools to remove the bad actors directly. Instead these changes aimed to retire the people doing the reporting along with the people they were reporting. There is a real chance good actors are now shut down for having spent their time calling out bad ones to improve the network. The changes themselves The response says SFDP was not reactive enough. On that, fair. The 1M SOL cap was badly outdated. But these changes do not trim fat. They cut into bone. The delegation cut: The validators who survive the retirement see their SFDP delegation drop from roughly 85k SOL to 55k. A 30k- 35k SOL cut is a severe hit for any independent validator, from any source. For many it is the difference between profitable and not. The twelve-month retirement: An automatic clock tells every operator in the program that they are a tenant with an eviction date. That is not how you retain the people you want. An annual review that renews contributors and removes the inactive would achieve the same accountability without the eviction notice. The 100-validator cap: If more than 100 operators meet the criteria, why turn them away? If fewer do, the cap is irrelevant. Either way it serves no purpose except to signal scarcity. Was any analysis done to arrive at 100? If so, can it be published? If the other changes are implemented, the program likely lands near that number on its own without needing to announce a ceiling. The timing: Breakpoint and Block Zero are days away. Independent validators who booked travel to attend are now being told they are no longer viable. Some even sponsored events. Some have cancelled their tickets. Some have already turned their validators off. Block Zero in particular is organized, hosted, attended, and sponsored by many of the validators now on the chopping block. At least one sponsor has said publicly that they would have reconsidered their sponsorship had they known they would be losing SFDP stake. These are exactly the people the network should be fighting to keep, and the Foundation chose the week before its own conference to tell them to leave. The communication problem underneath all of it Step back and look at the sequence. The announcement said: you are out, tranches start now. Nothing about rewarding contributors. Nothing about Alpenglow. Nothing about where the stake goes. After the community pushed back, the Foundation said contributors would be rewarded but could not say how. If the plan all along was to support contributors, that should have been the headline, with criteria attached. Instead the people affected have learned about their own future one X reply at a time days later. That is the part that has done the most damage, and it is the easiest to fix. In conclusion We are not asking the Foundation to reverse course. We are asking it to sequence this correctly. Our asks, updated from the original letter: ◆ Pause the weekly retirement tranches until communications around not throwing contributors aside is improved. ◆ The Foundation has now said Alpenglow community cluster participants should be rewarded. Additional details on this would be great because as it stands, these people are going to not be profitable soon. ◆ Publish the analysis behind the 300k, 100-validator, twelve-month, and 55k figures. ◆ Publish the impact assessment on the current validator set, including how roughly 75% of members came to be deemed non-contributing. ◆ Open a formal consultation with current SFDP members, with a stated process and timeline. ◆ Publish objective criteria and an application process for contribution-based delegation, to run alongside discretionary delegation. ◆ Answer the first-principles question: Why does the SFDP strictly value a new validator over a contributor who has honestly operated for years to improve the Solana network? The Foundation says these changes are about getting to what comes next. We agree that is the goal. What should not come next is booting Solana's contributors to the curb. L33

- double disinflation - VAT to zero - some % of SFDP stake to LSTs with contributor/participation focus - keep new SFDP rules/program for new operators and institutional direction empower contributors to delegate to peers, lower break even, increase asset price and adoption