gm
@solana is packed with insanely sharp builders! 🧠
that’s exactly why launching Hera v3 here first keeps feeling like the right call. the amount of work happening around liquidity, math, infra, tooling, blockspace, and execution is honestly different. we’ve taken a lot of inspiration from that energy while improving our own core around liquidity topology, agent-trained pair entropy, and deeper pathfinding.
@helius doesn’t stop shipping. even when things already feel good, they keep building side quests that make life better for both devs and users.
@jito_labs keeps pushing the execution and blockspace side forward, and that matters a lot when you’re building serious routing infra.
@JupiterExchange and @dflow are in the same arena as us, and respect where it’s due. they keep pushing for better trading, better output, and better market structure.
teams like @anza_xyz, @blueshift and the team at @solana_devs keep making the builder base stronger every day, constantly supporting builders and helping the ecosystem move forward.
and founders like @toly are always close to the community, reading the room, staying sharp, and thinking clearly about what comes next.
honestly, we’re enjoying this. now we just want to finish the execution layer, fire up the engine, and bring Hera to mainnet.
full throttle.
next stop: mainnet 🌝
development progress update ☑️
pipeline items close to done 👇
▪️ finishing the flow for integrated but not yet connected MMs
▪️ bringing new liquidity sources online
▪️ fixing two minor anomalies found in the fetching modules
▪️ improving the math for cases where the same source is used more than twice in routing
▪️ shipping a few small UI updates
major work still in progress 👇
▫️ programs + tx design + fallback flow
▫️ targeting optimal execution under v0 constraints
▫️ preparing the layer for the potential v1 transaction architecture
▫️ pushing route and call optimization even further
▫️ reducing centralization tradeoff risk and designing alternative relayer strategies
also in motion - side developments👇
▫️ infra improvements across the full VM stack
▫️ building clawbot, externally trained but powered by Hera data, for isolated review and improvement flows across all versions
▫️ moving propAMM strategy toward a fully verified model
▫️ redesigning the interchain architecture at a much broader level
no pain, no gain 😏
quick thoughts on the situation. 🤔👇
first off, no direct risk to $HERA here. we can confirm that both from the LayerZero team’s own statement and from our own post-mortem review. neither the token nor the omnichain staking portal uses OFT v2, both are on the v1, so isolated from the DVN side of this issue.
also worth saying out loud: the reason these exploits have started ramping this hard lately is AI. even an average dev can now get access to near-expert level tooling, and once that lands in the hands of talented but malicious actors, it becomes a real security problem for both protocols and users. odds are while you’re reading this, someone somewhere is already poking at infra, contracts, or validator setups with AI help, just looking for one crack.
and beyond direct technical attacks, the scariest DPRK vector is still social engineering. let’s be real about that. they pick targets, sometimes hundreds at once, and try to get close to the team or whoever sits near the keys. they show up polished, look capable, act patient, make themselves seem useful and trustworthy. they’re in no rush. they can do legit good work for weeks or months if that’s what it takes. then once they identify the right person and get close enough to start sharing files or workflows, they deliver the payload. after that, you know how the story goes. the scary part is how organized it all looks. either they already have playbooks ready for every type of protocol, or their internal coordination is insanely sharp. probably both. even if critical wallet material is kept outside normal access paths, teams will still need to touch it at some point, and these guys are willing to wait as long as it takes. that’s the reality. there’s basically no room for even one distracted moment.
looking at the recent LayerZero / KelpDAO exploit specifically, it gets even more uncomfortable. once you zoom out and look at the full picture, target chain selection, DVN analysis, chain-specific dynamics, RPC pressure, trying to influence validator behavior, and even the first moves right after the exploit, it really forces you to think about how serious this is getting.
some people, especially around Thor guys, have been saying wrapped asset arch is fundamentally dangerous. honestly, hard to say they’re fully wrong. but at the same time, if you’re trying to build innovative DeFi systems in current market conditions, you don’t get to just opt out of complexity. you have to find workable solutions.
as a team that has been building around LayerZero from the early days, we can say this clearly: the protocol itself is hard not to respect. DVNs are powerful if they’re designed properly. but if a protocol chooses custom configurations without fully understanding the risk surface, that part is on the protocol side, same story as OApp config. still, no doubt the LZ team will take major lessons from this. something like a stricter minimum security baseline wouldn’t have been a bad thing at all.
either way guys, being on the builder side is hard. being on the investor side is hard too. in moments like this, instead of instantly trying to dump all blame on one side, it makes more sense to focus on what can actually be improved.
wishing both @LayerZero_Core@KelpDAO@unichain calm minds and strength. hope things stabilize soon and you all recover mentally from this. also respect to @_SEAL_Org for stepping in and pushing on the solution side.
have a good week everyone. 🙏
this week we’re focused on smoothing out instability across a few propAMM integrations, bringing new liquidity sources online, and pushing an updated build with refreshed routing logic. on the UI side, 3 separate patches have already been shipped.
for the most critical part of the @solana rollout, we’ve started working on CU/instruction optimization and using ALT. we’re building an async decision layer that can keep route expansion within acc limits. at the same time, Jito bundles and deeper v0 tx flow upgrades are now in the pipeline, with research and experimental testing already underway.
lastly, it looks like we’ll be using @helius sender internally as well, still in testing for now.
still shipping, still deep in the trenches.
After a long wait,
Hera V3 beta is finally live on @solana 🔔
During beta, you can jump into the interface, check routes and quotes, and feed us with your feedback. Execution will stay disabled for now, and both the pathfinder and the UI will keep getting better every single day.
Bad routes, broken stuff, weird behavior, or anything you think should be improved, all of it is valuable.
Take a look now: https://t.co/KztpXoRu1Q ✅