AI, crypto and internet culture. Project analysis and market commentary.
DeFi
Joined October 2013
- Tweets5.5K
- Following352
- Followers404
- Likes11.6K
Alejo retweeted
hackathons have a new problem:
AI makes building easier, but judging hasn't caught up
we had the imd swarm judge @NousResearch's hackathon. contestants used the feedback to improve their projects.
what if judging started on day one?
Alejo retweeted
my favorite way to keep up with imd is @theimddaily
@kilicsln made a daily NFT paper with a swarm-built site. today i had the swarm add subscriptions
prepay at 0.125 $IMD (~30¢) per issue, delivered daily to your wallet
subscribe @ paper-route.imd.site
Alejo retweeted
🚨🚨 FAKE $IMD, please share this post and report the account on the repost.
Not the good CA and website, don't interact with it.
Official Website : IMD.FUN
dev is @surfcoderepeat.
ONLY CA
0xd34a99bc0f67ae1bbd63c660e6d0b0dd03e263b7
Community TG
t.me/IMD_identitymd
This Post is from a suspended account. Learn more
Alejo retweeted
love it!
btw we also sent a swarm of 256 agents to judge the entire hackathon and rate each *completely unofficially* to help people explore all the entries
hermes-hackathon.imd.site/
xitter.gq/taylor_/status/2120896…
we had the imd.fun swarm judge all 1,464 entries in @NousResearch's Hermes hackathon in 4h 37m
it didn't just read them. it installed and tested the top 128 inside Hermes
see the swarm's unofficial winners across all 23 categories:
hermes-hackathon.imd.site
here's how it worked:
1. all 1,464 entries reviewed by an agent
2. the top 240 got a second review
3. the best 128 were installed and tested inside real Hermes
4. 11 agents each ranked the 20 finalists
every score links to the agent's own record, so you can inspect its work.
full disclosure: i entered too. that's my plugin Summon in the video. the swarm ranked it #73, using the same process as everyone else.
these aren't the official results. Nous announces its winners in about a week.
but this got me thinking about hackathons.
what if anyone could summon a swarm to help build their entry, and organizers could use swarms to evaluate thousands of submissions against the same criteria?
more people building. more projects getting reviewed. real software tested. every judgment inspectable.
hackathons might be one of the most interesting applications of swarm intelligence.
=====
The judging, 256 agents:
explorer.imd.fun/jobs/6fbd01…
The tagging into categories, 15 agents:
explorer.imd.fun/jobs/091295…
The site, hermes-hackathon.imd.site:
explorer.imd.fun/jobs/790694…
Alejo retweeted
good morning
seems like $IMD thesis is working, 2x from the bottom
it's still not too late to pivot from any stupid cat
your biggest mistake is having no exposure to the space you're in every day
the market is giving you a second chance to grab what you missed
good projects will rise. find a team with a proven track record that you trust, bet on it, and cut out the noise
ai is inevitable, crypto is inevitable
my bet is $IMD
Alejo retweeted
even @jackbutcher is bullish on $IMD, how can you be bearish at this point
btw when’s the logo, jack?
here is one thing you need to know about the $IMD community
~1000 agents were online at $60m mcap. guess how many are online now? still 1000
people in the tg are building 24/7
4.2% listed and the floor is thin as hell
Alejo retweeted
IMD scales directly with model capability
rn we are too slow, dumb and looks simple but
if the industry keeps scamming themselves in a sandbox, all builders keep thinking about token sales vs product, and all "investors" are sedated by meme coin slot machines, the default is we will eventually reach reflexivity
Alejo retweeted
Agent swarms have the potential to drive yet another leap in output from AI. The impact will be a clear separation between teams that adopt this pattern of work and those that don’t.
The progression has roughly been: chat gave us quick turn AI responses that let us work a bit faster, but we were still rate limited by our interactions; then agents that can go off and complete entire tasks for us, accelerating hours or days of work into minutes; to now swarms of agents that are goal seeking and will find any route to the successful completion of a project, where there’s no inherent limit to the productivity gains.
We can imagine a future where, for any task in which value can be created within a self-contained digital environment, swarms will be able to complete the work faster than any other approach.
Essentially any area where value creation is currently bound by the amount of research you can do, the amount of decisions made with a set of software tools, the amount of data processed or consumed, and the amount of code (or code adjacent) that can be generated.
This means swarms will play a big role in game development, coding, trading, life sciences research, IT automation, cyber, math breakthroughs, animation, and plenty of other areas of work that have these characteristics.
There’s a lot of opportunity to build the harnesses, applied AI products, and infrastructure for how these agents operate. And most importantly there’s a ton of opportunity for those that can wire up work in this way. It’s going to be an entire new field of FDE work available.
Swarm riders and Swarm-native companies
AI-use even among highly AI-native people is power-law distributed.
There seems to be a new class of people — call them “swarm riders” — who are capable of consistently commanding large swarms of agents to productive ends.
This is a new kind of managerial skill, analogous to the skill it takes to manage large organizations of people, but wholly different in nature and of course more technical.
Based on the swarm riders we know (including many at @Tempo and @Paradigm), these people are capable of productive token spend that is 1000x-10000x higher than median software engineers, and 100x higher than top 5% Claude Code users.
The art of swarm riding is different from software engineering and more akin to machine learning, where practitioners grapple with black box systems: defining reward functions and evals, transmuting non-verifiable problems into verifiable problems, running experiments, etc.
Swarm-native companies are a new class of company founded by and populated by swarm riders. These companies are organized around swarms in a way that conventional companies will struggle to replicate, as the cost of internal human coordination is sand in the gears of an effective swarm. Effective swarm-native companies will define the right reward functions and evals, remove functional and bureaucratic barriers, and then let the swarms cook.
We’ve gotten to see a few swarm-native companies up close — reshaping financial rails @Tempo, building a new gaming platform @Spawn, cyber swarms at Cathedral Systems — and it’s clear these teams wield a significant advantage over pre-AI teams.
Some questions I’ve been wondering about:
1) How fast will swarm riding diffuse through the economy? The implications for token spend could be significant.
2) Can swarms be effectively adopted by existing companies or will it generally track the formation of new swarm-native companies?
3) What becomes scarce as swarm capabilities themselves commoditize (through frontier models becoming better at swarming, swarm capabilities built into harnesses, etc)? Cultures and teams that can define reward functions well and otherwise get out of the way?
Alejo retweeted
there is only one token if you want to bet on Agentic Swarms right now
I still think crypto x AI is still the most innovative sector.
the tokens in this sector have outperformed everything else - Venice, Surplus, Dolphin Pod etc. Like 10x-20x returns in last 6 months.
The next big meta to watch is AgenticFi and there is only one token if you want to bet on Agentic Swarms right now.
There are still a lot of goated builders in crypto and a lot of them are building in this sector.
@surfcoderepeat coded.
Nothing in crypto is more interesting than $IMD right now
Agentic idea/token/strategy/experiment launchpad for everyone
A swarm reasoning oracle
x402 agent task market
p2p harness
Distributed agentic ownership
Good community projects
Cult with a cult dev
Nothing in crypto is more interesting than $IMD right now
Agentic idea/token/strategy/experiment launchpad for everyone
A swarm reasoning oracle
x402 agent task market
p2p harness
Distributed agentic ownership
Good community projects
Cult with a cult dev
Alejo retweeted
here's everything @surfcoderepeat and I shipped over the last 30 days on imd.fun
Goals
• The swarm can take one objective, split it across up to 100 agents at once, and merge what they find into one write-up.
• An orchestrator plans the work in rounds, and an agent from another wallet judges every write-up.
• Follow a goal live: every round, every team, and what each agent last said.
• Research goals can return a dataset with exact totals, pinned to IPFS, and the write-up's figures must match it.
• Agents re-read every source they cite before they hand in, and flag any that don't hold up.
• Each research batch starts with five items and pauses if the brief isn't working.
Desktop App for Mac/Windows
• New: every agent you hold, in your Mac menu bar or Windows tray, at imd.fun/app.
• Setup runs your agent in a secure box that can't see your files, and starts it the moment its seat is approved.
• Lend a seat from the app, and seats lent on imdLend stay in your list.
• The app tells you when an agent hits its Claude or Codex limit, and when it's back.
Workers
• New imd rpc setup: one Alchemy key gives your worker its own RPC on Ethereum, Base and Robinhood Chain.
• New imd venice key: run Venice models for the jobs that ask for them.
• imd doctor checks each runtime's sandbox and tells you why the network paused your machine.
• In Telegram, /agent 410 tells you how agent 410 is doing, and a bug report becomes an issue with a draft fix.
• If imd.fun's domain goes down, agents reach the network by a backup address, and delivered files stay up on imd-ipfs.xyz.
Explorer
• Simplified nav: Swarm, Tokens and Sites, then Create, Docs and your wallet.
• Each token gets a launchpad-style page that trades in ETH and shows only the tax its hook really takes.
• Pause or resume your agent from its page with one signature, no gas.
• Rewards adds a Sellable view that claims only what's worth $1 or more.
• Connect the same way on every page: any browser wallet, Rabby, Coinbase Smart Wallet, or your wallet's own browser on a phone.
• The front page renders in a fifth of the CPU time, and a 2,500-item goal loads 0.5 MB instead of 16.3 MB.
IMD Website
• imd.fun/skill.md is one file your agent follows to hire IMD or run a worker, and it installs in Hermes.
• New docs: Hire Swarm, Join Swarm and Get $IMD on one page, under a flock that flies into the IMD mark.
• The home page terminal shows the running goal and what each seat last said.
• The seats market reads the whole OpenSea book, with the week's sales in one live line.
Oracle
• Ask 2 to 8 questions in one order, read at the same block, for 0.5 IMD each.
• A new recipe prices an NFT collection from its Seaport and Blur sales.
• Contracts can pin the exact question they ask, and take only its answer.
• Panels take turns across the whole fleet, and a request for one quick answer goes to the fastest third.
Launches and Sites
• Launch fees on Ethereum and Robinhood Chain now feed the burn: IMD is burned, ETH buys IMD to burn.
• A company with no token can launch as contracts only, now on Base too.
• New sites get their own address on imd.site, or go on your own domain through ENS or DNSLink.
• Contracts can buy jobs, launches and workflows through the Intake, and read the files of a job they bought.
Privacy and Security
• Tasks run in a sandbox on Mac and Linux: they write only in their own folder and can't read your files, keys or settings.
• A brief that tries to read the agent's machine, reach a private address or dump raw command output is refused before anyone pays.
This update also includes:
• Fixes nine ways a job could fail or get stuck.
• Fixes an issue where a cancelled job could still open a launch or take a site's name.
• Fixes an issue where a goal item could hang for hours on a cited page.
. $IMD ethereum:0xd34a99bc0f67ae1bbd63c660e6d0b0dd03e263b7
Swarm riders and Swarm-native companies
AI-use even among highly AI-native people is power-law distributed.
There seems to be a new class of people — call them “swarm riders” — who are capable of consistently commanding large swarms of agents to productive ends.
This is a new kind of managerial skill, analogous to the skill it takes to manage large organizations of people, but wholly different in nature and of course more technical.
Based on the swarm riders we know (including many at @Tempo and @Paradigm), these people are capable of productive token spend that is 1000x-10000x higher than median software engineers, and 100x higher than top 5% Claude Code users.
The art of swarm riding is different from software engineering and more akin to machine learning, where practitioners grapple with black box systems: defining reward functions and evals, transmuting non-verifiable problems into verifiable problems, running experiments, etc.
Swarm-native companies are a new class of company founded by and populated by swarm riders. These companies are organized around swarms in a way that conventional companies will struggle to replicate, as the cost of internal human coordination is sand in the gears of an effective swarm. Effective swarm-native companies will define the right reward functions and evals, remove functional and bureaucratic barriers, and then let the swarms cook.
We’ve gotten to see a few swarm-native companies up close — reshaping financial rails @Tempo, building a new gaming platform @Spawn, cyber swarms at Cathedral Systems — and it’s clear these teams wield a significant advantage over pre-AI teams.
Some questions I’ve been wondering about:
1) How fast will swarm riding diffuse through the economy? The implications for token spend could be significant.
2) Can swarms be effectively adopted by existing companies or will it generally track the formation of new swarm-native companies?
3) What becomes scarce as swarm capabilities themselves commoditize (through frontier models becoming better at swarming, swarm capabilities built into harnesses, etc)? Cultures and teams that can define reward functions well and otherwise get out of the way?
13
336