Well, you’re confused about which comes first. End users are the last to come. Look at Bitcoin, look at Ethereum.
I’ve always said that the miners come first. Along with them come the speculators. App Builders are next, and end users won’t even know they’re using one of tens of hundreds of interconnected bridged Autonomi Networks, each with a different underlying storage-paying token…
Check out the Dev Forum
we need to finally release a MVP … to see if the basic concepts hold to their hopes/promises in the wild - no point in releasing “the perfect product” after 18 years … it will be flawed if not tested under real world conditions in the process … there is no such thing as the perfect product … and for sure you can’t know how close you are if you don’t test …
oh - just a bit more than 2 petabyte on average farmer size … I think I saw insane numbers before - but forgot about it again - super impressive indeed xD
Monopoly money sizing. How many of the beta nodes have 500 nodes with 2TB backing them. Claim of 16TB yet really only 2TB
You are confusing what impressive means.
When something can be bruteforced it’s not impressive - it’s just money as a multiplier.
The impressive thing is when I first reset my Autonomi nodes and after 5 minutes I was winning nanos again.
Of course, for this to impress you, you must have lost money trying to be a SIA/Storj miner - every reset is death and a waste of months of work.
The elegant beauty of the Autonomi Network’s ever-moving river of data is something truly impressive that will bring millions of farmers into the game and is the first step towards the ideal set by the Safe Network…
Check out the Dev Forum
I’m sure a 4MB chunk size performs better than 512KB when uploading or downloading large files. But I’m uneasy about the effect on cost for small uploads. A lot of files people will want to upload will be much smaller. And the price to upload them will be driven by the minimum size. It will make the usage for small files less attractive. Has this been considered?
Yes that’s certainly a factor as well.
But so too is a 70% improvement in performance, and reduced gas too
gas, gas… if we only had the real Native Token… But how is that practical to upload websites to Autonomi now? There’s a ton of small files, html, css, pictures, etc. much smaller than 4MB. Wouldn’t this step be against such use case?
Do we really want any erc20 related stuff influence the design of the most basic structure of the network?.. If wenn do then @happybeing is right and this erc20 stuff is detracting us from our goals… At least the ones I am on board for…
I thought this is meant to be just an small episode of erc20 training wheels because the native currency is not ready yet… Bundling payments, optimising chunk size for it, (delaying the api because the initial wasm concept and connecting to the erc20 stuff doest play that well together…? Just a shout in the blue because the api is still missing)…
Performance in which use case…? For sure not browsing autonomi
But I guess the logical step from app builders will be 1. Bundling websites to monolithic blobs instead of many small files (dedup will then no longer be a things for those) and 2. Apps will be so the same for their logic… So a autonomi dropbox will try to squeeze as many files into larger blobs as possible to reduce cost on usage…
No, you can still upload things smaller than 4MB of course
In response to the OP.
Seeing exactly the same characteristics, node’s mostly under 1000 records, and under 3GB per node disk usage - not even close to 32GB per node 
Plus points, Home-Node seem to be earning, which is really positive - although there is an earnings bias I’m seeing towards nodes being setup with port forwarding - will be interested to analyse the EVM proposal, to see if that improves this situation - there should be an even playing ground.
As for this testnet, I have to say I’m not convinced we are testing Apples, with Apples to draw a conclusion on chunk size ! 4MB’s to save GAS fees makes me sad to see 
While I appreciate my netflow stats aren’t accurate (and node numbers aren’t as churn is hard to plot with port re-use), they paint a picture I’ve had concern over since these beta’s started a few months ago - we seem to be very VPS heavy, maybe from generous people supporting the project - but maybe the fluid nature of the rewards is also a reason - at any point, these machines and nodes are biasing the tests - Even the foundation nodes are predominately droplets ?
It would be useful to help our understanding, to get more insight into how 4MB Chunk size given the VPS bias - #edit - my concern is it feels like we are excluding things like RPI / Synology / QNAP / and other low power home devices - why are we bothering with Windows, and home network mode - if these devices don’t represent more than 5% of network… Would be really interesting to see a network with home machines only…
Also, the Disco Rank bot - ~ 500 unique participants - yet Wave1+2+3 should be close to 2000 unique participants - doesn’t that variance cause any concern ? have we just lost / disengaged with 1500 people 
Jad
Well, 4MB chunks seemed to beat me the performance of any other chunks size tested in every metric other than node memory usage.
Well yes, likely including that too. But we also have other strategies to even further improve their through caching layers etc.
Just to be clear, that’s not been part of the decision at all, just a byproduct of it.
Even if we were going native token from the off, we would still have arrived at this result.
We do still have further internal tests running, and then ofc the public test this week, so it is still subject to the results of all those too.
And no, we are at about 1500 unique participants for this phase.
nd we know that some in the waves juist signed up but didn’t ever run nodes; so I think we are similar but steadily growing with the current net.
Hi Jim, thanks for taking the time to respond - I do appreciate it…
Ok I won’t panic 
I would just like to not see the tail wagging the dog when it comes to compromising network scalability / usage against the temporary inconvenience of Gas…
I’ll leave it there and get back in my box…
Have 1000 just not earned any nanos then Jim?
I don’t have theee most up to date figs, but that’s in the ballpark yeah
What is the evidence for this, please?
Please don’t tell me its IP addresses, cos I account for 2, sometimes 3 of these. Other individuals easily account for 10+ themselves, so please forgive me if I take the 1500 figure witha large dose of salt.
TBPH, I’d put it at maybe 500 individuals and thats being generous.
Risky strategy, but just for a few days, try running a TestNet without rewards and lets see how many turn up…
It might not be a lot but we’d soon see how much VPS use affects results.
I have some machines running just 4 nodes and all have earned, it’s worrying that 2/3rds of participants have yet to earn a nano don’t you think?
