As we approach the token generation event (TGE), announcements from the team have been coming thick and fast. There’s not too much to add in our general summary this week, so we’ll just round them all up in one place.
Network size
First a glance at our network size chart shows that we peaked at something like 290,000 nodes and have (at time of writing) settled on a reasonably steady state of about 280,000. This equates to roughly 1.7 PB of user data storage capacity in 8.5 PB of raw storage capacity. @shu has run a comparison: Sia Network has 86,000 nodes and roughly 8.67 PB of user data storage capacity. Storj Network has 26,000 nodes with roughly 112 PB of user capacity.
Latest release
If you haven’t done so already, please upgrade to the latest releases: antnode v0.3.5 and node-launchpad v0.5.3. Community members are already reporting leaner operations and better stability - so what are you waiting for?
Smart contract upgrade
We are planning very shortly to upgrade the smart contract for chunk upload pricing. The new contract will adjust upload prices dynamically based on the network’s utilisation, resulting in dramatically higher upload prices and rewards.
Airdrop
We also announced details about the upcoming MAID/eMAID to ANT airdrop. If you’re a MAID or EMAID holder, be sure to read this post.
Winner winner chicken dinner!
Social media prize goes to @LJlang. Yay! 
Dave news
Dave is up and about and the more feedback we get from you folks the better. So please give him a whirl and tell us your findings.
We know supporting only mobile wallets is not ideal; this is due to technical limitations with desktop support of the available wallets, and the WalletConnect protocol which we are using in order to provide the broadest coverage for wallets that we can. This doesn’t mean we aren’t looking at ways around it, just that this is why we released with Mobile only wallets.
We’re working on addressing all your feedback, and we greatly appreciate those folks who have provided it. We continue to work on Dave, but at this moment the network is the top priority, so some resources have been shifted there for the moment.
General progress
In alphabetical order, cos why not?
@anselme undertook a massive clean-up of the autonomi file structure, as well as separating code into clean modules for each low- and high-level datatype. This should make it simpler for us to build on. He also wrapped up the register removal, and cleaned up areas of the code where ‘dead’ types like Spend, transaction and auditor were still present.
@bzee has removed mDNS from the codebase. mDNS is a mechanism for peers on a small LAN to discover other peers. We were using it to bootstrap local test networks for convenience. By stripping it out, we now rely on the same bootstrapping mechanism as for a non-test network. This means we have fewer differences between our test and production networks, which has already uncovered a few optimisations for our bootstrapping process.
@DannyBeck is the main Dave man (Dave being the client GUI). He’s been fielding comments from the community and making changes accordingly. As mentioned above, we’re aware of a number of bugs and UI suggestions (getting this feedback ASAP was why we released Dave when we did). Keep the suggestions coming!
@dirvine has been looking with @shu, @roland and @qi_ma into the issue of whales riskily overprovisioning servers in the early stages of the network, aiming to go beyond the rather crude prevention mechanism we currently have in place. @qi_ma has removed this mechanism, and we’ll see how it goes. 
Ermine worked on replacing RPC with a metrics server for working with dynamic string data.
@jimcollinson has been working on a new documents site on GitBook. More on that soon.
Lajos wrote the airdrop script, and is working on unit tests for the Revenue Split, Leaderboard, App Registry smart contracts.
@mick.vandijke debugged an upload/download problem that’s been giving us gip, and fixed an issue where on zero collected quotes, the client would consider the chunk address to be already paid.
@qi_ma raised a PR to prevent free data uploads through replication, and one to avoid the CPU refreshing too frequently, which was interfering with metrics collection
@roland raised a PR to prevent harmless incoming connection errors from affecting the external address score, and one to introduce a relay reservation score to improve home node connectivity.
@shu worked on identifying improvements revolving resource consumption when gathering metrics, suggesting further improvements based on his initial analysis, as well as working on upscaling and DCUTR tests
Thanks to @happybeing for his suggestions for some changes to the Archive API in order to improve utility, privacy and reduce the likelihood of breaking changes.
WEB10
whatever the value in that should be …