We're working on deploying the modicum implementation to several chains to benchmark the performance of the current contracts and code on them and identify potential bugs.
The research into multi-agent reinforcement learning continues, as well as the Lilypad Whitepaper commencing!
HackFS Winners were also announced last week! Check out their awesome projects!
⚒️ Lilypad Engineering Update
Escape from the rabbit hole
Much of last week was spent trying to figure out why the changes made to attempt to make IPC work had broken integration with hardhat.
Painstakingly testing each commit in turn, and juggling various configuration issues and other workarounds, we eventually narrowed it down to the way we adjusted filter handling to work with IPC, which doesn't provide the entire block history. We came up with a workaround and then were able to confirm that the latest code did indeed work with hardhat, so could convince ourselves that we hadn't broken anything making the changes to work with IPC, which unblocked the following item...
Minimal reproducer for IPC
We were then able to add code to the new EthereumClient to log to a set of pickle files exactly the web3.py interactions it was making which were still failing against IPC. This unlocked writing a minimal reproducer that the IPC team will be able to run, which looks like this when run against hardhat:
But which looks like this run against IPC:
Sadface, it hangs forever waiting for the transaction confirmation on the function call.
We're working closely with the IPC team to run down this bug as quickly as possible! So far, they think it has to do with calling functions that emit events...
Better developer experience for contributors
We've been working on making the MODICUM repo easier for external contributors to work with, in particular, we've made a lot of fixes to the README:
🎓 Lilypad Research Update
Levi has been continuing his research into multi-agent reinforcement learning over the past week.
There have been some exciting updates in the works regarding anti-cheating mechanisms and pricing models in permissionless and trustless distributed computing environments. These improvements should substantially reduce several thorny issues that prior protocols have not addressed.
Work on autonomous agents is making steady progress. Expect to see updates around the time that testnet launches.
The Lilypad Whitepaper also commenced this week. 🎉
🌟 Lilypad "All the Things" Update
This week, the winners of the HackFS Hackathon event were announced!
So many great projects were submitted to the Bacalhau/Lilypad track of HackFS that we even ended up giving out an extra prize (and we wish we could have awarded more!)
These winners join the ranks of past HackFS winning projects such as Huddle01, Spheron & Lighthouse Storage (who were all sponsors this year!), as well as Secured Finance, Valist, EthSign and many more successful startups in the web3 ecosystem.
We're looking forward to seeing where these projects go next!
🏆 HackFS Winners Announced!
It was amazing to see an influx of ideas and innovations in both the Slack channels and through the final submissions and we hope to see more of these projects continuing to build in the future!
See the full final ceremony here:
As mentioned we had so many high-quality applications for Lilypad and Bacalhau in HackFS, that we ended up giving out 6 prizes rather than 5!
The 6 winners of the $2000 each were:
The DefiKicks team put an impressive amount of engineering work and skill into building out a decentralized, community-governed Data DAO on Filecoin that democratizes DeFi data aggregation and TVL calculations.
Bacalhau plays a pivotal role in DefiKicks by executing off-chain jobs, such as the computation of inflationary rewards and the tallying of off-chain votes, ensuring a smooth and efficient voting and rewards process.
Daggle made Bacalhau easy by adding an awesome UX and Marketplace for finding Bacalhau jobs. The prize was awarded for the potential high impact Daggle could have for common usage of Bacalhau / Lilypad in the future.
Tentai made AI easy by creating a "Decentralized Huggingface" with Bacalhau. It's what we've all been looking for lately!
InferAI is the AI UI I've always wanted for Bacalhau! Thanks team - we hope you keep building!
Pensieve impressed us with both their techincally savvy submission that used Bacalhau for image processing and their grand idea for creating not just personal memories but a full documentation of Homo Sapiens or History of the Apes as they call it.
🏅 DYield Collector
DYield Collector is a proof of concept of decentralized yield data aggregator. The team used their own custom script on Bacalhau to compute the best yield data to get the best APY value of the reserve to send to the user with a push notification.
We want to also give a special shout-out to the following teams, who I wish we also could have included in the finalist list and want to encourage to keep going!
- We loved the idea of an ML Data collective!
- This concept reminds us of this amazing talk by Peter Wang. Decentralised AI in the new era is going to be ever more important
- Special shout out to Kazeem here for the PR adding his work to the examples docs on Lilypad enabling other folks to use his solution as well!
- Making chatGPT your front end for running Bacalhau jobs? Yes Please! UX ftw!
- A fantastic idea for crowd-sourcing images for training medical analysis tools.
🤓 CODwg is back!
It's revenge of the nerds, wait... I mean, the COD working group regular catchups will recommence from the 7th of July next month!
You can catch up on the backlog of videos from groups building web3 compute over data solutions in the Youtube playlist below, including Taubyte, New Atlantis, Polybase, Phala, Fission, Charity Engine and many more!
Special shout out to Wes Floyd for building and hosting this group over the past 6 months, building strong networks and community collaboration in the space.
Details next week on guests and where to tune in!
🔮 What's Next?
We'll continue working on getting a modicum testnet up for mid-July. This includes benchmarking results of the current code across several chain networks and working on the whitepaper.
☎️ Contact Us
🗞️ Subscribe to the newsletter: https://lilypad.tech/newsletter
Did you find this article valuable?
Support Lilypad Network by becoming a sponsor. Any amount is appreciated!