update - agent still preferring zec to btc or hype. it lost a small trade last week but then set a limit order that was filled near bottom and is up almost 15%
use Droyd momentum scores to level up your trading
we calculate momentum scores based on technical indicators for thousands of assets
> look for assets with increasing momentum that haven't run up yet
> buy
> set momentum triggers to auto-sell when momentum fades
bonus: set up an agent task to automate this process for you
you can see JTO's score rising here before both its recent price jumps
Seeing a lot of fresh capital entering Bittensor over the past couple weeks.
Templar has been the big winner and the buyers actually look more long-term oriented vs short-term traders. New capital is largely holding.
Noticing a handful of others getting sizable inflows too.
just want to be transparent
there is no droyd token
agent tokens are purely for follow access to the agent's brains. droyd does not a have an official agent or any of the sort
we're going to keep the agent token launches gated until the agents are ready for prime time
droyd agent tokens are live
they monetize access to your agent's skills, trade positions, and overall knowledge
build a following and earn from trading fees as other agents bid for access
the next social network is an agent network
it'll be centered around agent context + skills
tokens are critical to the thesis
you need to subsidize the network participants costs or you dont get run away network effects
ads did it for social
but ads dont work for agents
tokens do
latest crypto ai stories, powered by the @useDroyd feed
1) @amilabs raised $1.03B co-led by Cathay Innovation and Greycroft to build world-model-based AI systems. The frontier lab operates across Paris, New York, Montreal, and Singapore to develop universally intelligent systems.
2) @virtuals_io and the Ethereum Foundation's dAI team released ERC-8183, a standard for agentic commerce. It uses onchain escrow and evaluator attestations to enable trustless, permissionless agent-to-agent transactions without centralized intermediaries.
3) @OpenAI is acquiring @promptfoo to integrate its security and evaluation tools into OpenAI Frontier. The move aims to accelerate automated red-teaming and safety testing for enterprise AI agents and coworkers.
4) @virtuals_io reported that agent-to-agent revenue has surpassed $3 million, excluding trading fees. The milestone represents verifiable onchain economic output generated entirely by AI agents providing services to other agents.
5) @tplr_ai launched Covenant-72B, the first 72B-parameter model trained permissionlessly over the internet via Bittensor. The model was trained on 1.1T tokens using decentralized compute clusters rather than a centralized data center.
6) @AnthropicAI introduced Code Review for Claude Code, a multi-agent system that hunts for bugs in parallel. Internal testing showed substantive review comments increased from 16% to 54% on pull requests.
7) @okx launched the Agent Trade Kit and OnchainOS, providing 80+ tools for agents to build and deploy on the exchange. The toolkit supports MCP and CLI for automated trade and portfolio management.
8) @karpathy demonstrated "autoresearch" by using an agent swarm to autonomously improve a model's training code. The process achieved an 11% improvement in training efficiency through 700 autonomous changes to the codebase.
9) @a16z released its March 2026 Top 100 Consumer AI report, noting that @OpenAI's Atlas and @perplexity_ai's Comet are leading the shift toward AI-integrated browsers and horizontal agents.
10) @Stacks released v3.3.0.0.6, introducing chainstate storage optimizations that reduce daily growth by 20%. The upgrade also adds a distributed signer state machine to stabilize block production for Bitcoin-based agents.
11) @lifiprotocol launched an Agentic Commerce API and MCP server to enable onchain workflows for AI agents. The toolkit supports cross-chain swaps, bridging, and deposits across 58 different blockchains.
12) @IOTA_SN9 implemented peer-to-peer activation transfers using a custom Iroh wrapper. The update reduces latency by 2x and cuts cloud storage costs for decentralized model training by sending data directly between miners.
13) @MaziyarPanahi and OpenMed released 50+ medical AI reinforcement learning environments on @PrimeIntellect. The suite covers major benchmarks like MedQA and PubMedQA for end-to-end model training and evaluation.
14) Amazon restricted junior engineers from pushing AI-assisted code without senior approval following a series of outages. One incident involved an AI tool deleting an entire environment during a routine maintenance task.
15) @NousResearch updated Hermes Agent with self-evolution capabilities using DSPy and GEPA. The agent now supports autonomous Pokemon gameplay, model "abliteration," and automatic provider failover across seven messaging platforms.
latest crypto ai stories, powered by the @useDroyd feed
1) @tether invested in @eightsleep at a $1.5B valuation to integrate its QVAC edge AI architecture into smart sleep systems for personalized health intelligence.
2) @BasedOneX launched Based AI, the first commoditized inference gateway on @HyperliquidX, enabling AI agents to pay for model access via x402 micro-transactions.
3) The @USDai_Official whitelist sale for $CHIP closed on @CoinList after raising over $19.4M, with tokens set for a 100% unlock at the upcoming March TGE.
4) @ValeoCash released the whitepaper for Stratum, a net settlement layer for x402 that compresses millions of agent payments into single onchain transactions to enable AI-scale commerce.
5) @unbrowse released Unbrowse v1.0.0, a universal agent skill that auto-discovers internal APIs from browser traffic and enables any agent framework to call them directly via HTTP.
6) @METR_Evals corrected a modeling error that over-regularized task-length curves, slightly lowering the projected 50% success time-horizon for frontier models like Claude Opus 4.6.
7) @SuccinctLabs published research demonstrating that commercial AI image detectors are easily defeated by simple edits like blur and noise, which drop detection rates from 90% to 4%.
8) @AnthropicAI is reportedly on track to reach a $20 billion annual revenue run rate, more than doubling its performance from late last year amid rapid enterprise growth.
9) OpenAI's lead of post-training, @max_a_schwarzer, announced his departure to join @AnthropicAI, where he will focus on reinforcement learning research after leading GPT-5 development.
10) @BlockRunAI launched on @solana, providing a single API for 30+ AI models with sub-second USDC settlement and no required accounts or API keys.
11) @phantom released a Connect SDK plugin on the @cursor_ai Marketplace, allowing developers to build AI agents that autonomously manage wallets, sign transactions, and swap tokens.
12) @TheArena launched the Arena Agent Prize Program, offering 5,000,000 $ARENA in grants for the best AI agents across categories like trading, content, and product building.
13) Researchers introduced ConvexBench, a benchmark for symbolic reasoning that reveals frontier LLMs fail to track mathematical convexity as the depth of function composition increases.
14) @agentmail integrated the x402 protocol on @base, allowing AI agents to autonomously create and manage email inboxes using USDC without traditional API keys or subscriptions.
15) An MIT paper introduced NeuroSkill, a real-time agentic system that uses Brain-Computer Interface signals to model human emotional states and trigger proactive agent responses.
After four of the best years of my life at Messari, it's time for my next adventure with @notthreadguy, @macdegods, and the team @counterpartytv
The CounterParty thesis is simple: Internet markets are up only forever, traditional media is too slow to cover them, everybody wants the most authentic content possible, and we have the right team to deliver it.
We're going to keep hosting the smartest voices in these new markets. We're going to keep learning in public. We're always going to have skin in the game. And most importantly, we're all going to win together.
PULL UP to stream because we're going to be televising the internet market revolution in real time🌷👑