investment & research @ambergroup_io

Joined May 2022
151 Photos and videos
Pinned Tweet
29 Sep 2025
crypto bros flying across the world to Korea and Singapore so they can stand around at events like this
139
112
3,564
165,463
Jun 5
in light of all this ZEC discussion i just want to say one thing … nocktra
12
5
88
2,981
May 6
im writing a report on programmable privacy. i realized comparing projects is sort of dumb. for example, a nockapp on @nockchain will have private txs, but probably still wants FHE from @octra when users need to compute over each other's private data report soon
8
5
54
2,734
Apr 14
some crypto people faded prediction markets because "they arent really a blockchain thing" but you need to understand perps are a blockchain thing and perps will be wildly important to any prediction market achieving leverage
7
18
712
Apr 11
i don’t care if you don’t believe me but i believe this year some of my most important connections will be made on the @praxisnation mobile app
5
27
4,254
Apr 10
all the best teams eventually leave bittensor. learned this from nous leaving their subnet and wombo leaving their subnet. RIP the best narrative bittensor had in awhile
3
2
22
4,630
Apr 8
if a large crypto company ever gets funded in toronto please make a third space like monad did in nyc
7
1
27
2,339
Apr 8
>large crypto company >toronto prob never happening unfortunately
1
8
599
Apr 2
you always see many companies building "AI agents for finance" but very few are building the stack that makes them actually work at institutional scale that's what amber is doing
"In the AI agent economy, financial services will inhabit agent-native OS just as web apps rely on the cloud today. Crypto is the economic rail embedded in this architecture." @MichaelWuAmber at @EthCC 2026: "We aren't just adding AI; we're building the financial stack for what comes next." 🧵👇
1
11
700
Apr 2
Amber Group.
3
253
Mar 25
there is so much alpha in looking at a project’s dev branches and seeing what they are working on / where their head is at
12
465
Mar 25
meeting people in nyc it’s increasingly clear that there is a new generation of free thinkers who will run laps around the old generation imagine being a grey haired unc who is ideologically short hype and still doesn’t even know what bittensor subnets are because you are never online
2
11
433
Mar 15
looking back at this report its clear to me that proof of work coins, whether or not fully true, provide a perception of fairness in an industry plagued by max extractors. when the average person sees "mineable" they think Bitcoin. they think fair launch. they think nobody got a premine. that perception alone is worth something in a market where every new token launch is some variation of insiders dumping on retail. bringing out the history books with ERC-918: eips.ethereum.org/EIPS/eip-9… after the first ever ICO wave, many smart people of the time of the time realized that ICOs had largely cooked retail. the token deployer controls the entire supply, raises millions, and retail is left holding the bag hoping the team delivers. their solution was a mineable token standard. basically ERC-20s you could mine. an "Initial Mining Offering" instead of an ICO, where the deployer is "just another user" and tokens are distributed through proof of work, not through a presale allocation spreadsheet. their idea was largely correct. at a high level this would be a more fair way to launch tokens. but i've done research into how we could bring this back, and their implementation was stupid. you had to call the mine() function in the smart contract every single time you wanted to earn tokens. that means every solution required an on-chain transaction. at that point you are just exchanging gas fees for tokens. sure you accomplished the fairness goal but the work itself is meaningless. BUT SINCE 2018 WE HAVE EVOLVED. the work can now be useful. AI inference, model training, ZK proof generation, these are all workloads that actually demand real computing power. and we are starting to understand how to coordinate this computation in a distributed manner akin to Bitcoin, which was always the original vision behind something like ERC-918. i imagine we will find ways for this distributed computing power to be genuinely valuable. not just burning cycles to prove you burned cycles, but producing outputs that people actually pay for. i also imagine a world where speculating on a token inadvertently creates real demand for that token, because buying and holding it means someone is mining it, and mining it means compute is being contributed, and compute being contributed means the network is actually doing something. speculation and utility become the same loop. that's the version of ERC-918 that should have existed from the start.
Replying to @ambergroup_io
12/ Read the full deep dive → “Proof of Useful Work” Written by Amber Researcher @robcrt ambergroup.medium.com/proof-… If you are building something related to Proof of Useful Work—we want to talk. #PoW #ZK #AI #AgentFi #AmberGroup
3
2
26
1,218
Mar 15
pumpfun trenchers this weekend after discovering bittensor subnets
2
1
15
804
Mar 12
how it feels to watch a non crypto project's discord chat get invaded by 100 members of a spammy shitcoin DAO begging to claim fees now i know why normies hate us
11
395
Mar 8
prediction markets and equity perps prediction markets and equity perpsprediction markets and equity perpsprediction markets and equity perpsprediction markets and equity perpsprediction markets and equity perpsprediction markets and equity perps maybe ai this is crypto now
2
9
329
Mar 8
opened reddit and yep looking back i have bought the wrong illiquid speculative assets i should have went all in pokemon cards
2
7
512
Feb 6
this ethdenver side event just got canceled rip retail
2
18
1,077
Jan 27
im in sf and someone pitched me that he can make gold
7
22
868