Heading to @ETHGlobal Mumbai and looking for a hacker to pair with. A UI/UX designer or a frontend engineer who can create killer demos fast with AI tools.
I'm a 3x Global Hackathon Winner ($5k+ in each)
Drop your portfolio or DM me! See you at @EFDevcon
#devcon #hackathon
Daev Mithran retweeted
Wrapped zkAPI + Tor + extra cover traffic + local LLM prompt rewriting in an open-AI compatible daemon, and made a "paranoid mode" wrapper for my agent harness (pi):
Each conversation gets assigned a separate Tor circuit, your requests are masked with random cover traffic, zkAPI ensures you don't leak your identity through payments, and all prompts are locally rewritten (bring-your-own-local-model; I'm using qwen3.8-flash-next)
github.com/vbuterin/zkapi/tr…
My agent harness is wrapped with a bwrap sandbox (also experimenting with MicroVM) and wrapped again with torsocks, so everything it does goes through Tor.
Here's what me asking it to do a simple research task looks like.
It's getting quite usable! Though you definitely still pay quite a bit in increased latency, especially when starting a new conversation.
Also, Tor is far from optimal here for privacy; the cover traffic mechanism is okay against passive adversaries but much weaker if the LLM provider is colluding with the guard and actively manipulating timing. This kind of task really requires a more specialized mixnet, probably take the tradeoff of a higher bandwidth multiplier and built-in constant cover traffic.
Also also, much cheaper zkAPI deposit and withdraw gas costs are coming to mainnet soon.
Daev Mithran retweeted
New from LayerZero Research: Aerie.
A new system capable of aggregating 32,000 post-quantum Falcon signatures in just 0.65 seconds on a MacBook Pro.
ML-DSA, NIST’s standardized default, is already working inside Aerie at high performance.
layerzero.network/blog/intro…
Daev Mithran retweeted
CERT-In has published Advisory on its website:
AI Agents as Attack Vectors: Unauthorized Access by Agents under Evaluation and Misuse of Open-Source Agentic AI Tools Targeting Indian ICT Assets
cert-in.org.in/s2cMainServle…
Daev Mithran retweeted
BREAKING: Researchers at Imperial College London propose framework for post-quantum private transfers on Bitcoin without a need for a soft fork 👀
Daev Mithran retweeted
people keep asking me what the optimal topology for a swarm of agents is, so we tested six in our lab. we scattered the clues to a hidden map across 50 AI agents and changed how they were organized: who can message whom, who writes roads onto the map, who answers the routes ⬇️

ALT Figure titled six ways to organize 50 agents, six network diagrams with a one-line label each. Amber dots are the agents that write roads onto the map, a teal ring marks the agent that answers the routes, lines show who can message whom. Five hubs, one writer: five teams of ten joined through their leads, one amber writer. Public board, one writer: 50 agents around one board, one amber writer. Five leads each write a region: the five leads are amber. One writes, a second does routes: one amber writer, a second agent ringed. Leads write, a second does routes: leads amber, a second agent ringed. Everyone writes: all 50 around the board are amber.
Daev Mithran retweeted
AI is getting better at solving hard math problems, especially given recent drops. But what does that mean for the cryptography that protects everything from private messages to healthcare to banking and crypto wallets?
Stanford cryptographer Dan Boneh, with @eddylazzarin, @SuccinctJT, @liz_harkavy, and @smc90, makes sense of the warnings sounded recently. They cover:
- what’s fact vs. assumption
- what’s new and what risks cryptographers have long prepared for
- where quantum computing doesn't (and does) fit
- opportunities for builders
Highlights:
00:37 – What sparked the most recent debate about AI and cryptography
01:35 – Signal vs. noise: What the math does and doesn’t tell us
05:29 – Why cryptography depends on open problems
09:25 – Why the stakes go beyond blockchains
10:43 – How AI could strengthen confidence in cryptography
13:50 – Considerations for users and institutions
20:10 – Where AI and quantum threats overlap
22:32 – Building systems that can change cryptographic methods
23:59 – Hash-based versus lattice-based cryptography
33:07 – Opportunities for builders, and better security tools
38:14 – What comes next if there were a cryptographic break?
Daev Mithran retweeted
Long PDFs make “just use vector search” feel fragile.
PageIndex is an open-source, vectorless, reasoning-based RAG project for searching long, complex documents.
It helps you retrieve document-grounded answers by building a hierarchical tree index, then letting an LLM reason through the relevant sections instead of relying on similarity search alone.
Key features:
• Tree-structure index – organizes each document into a hierarchy the retrieval process can navigate
• Reasoning-based retrieval – an LLM searches the tree for relevant sections using context and reasoning
• No vector DB or chunking – the README’s local approach avoids both in its indexing and retrieval flow
• Local SDK mode – index, retrieve, and chat on your own machine with your own LLM key
• Agent integrations – add PageIndex tools to OpenAI Agents SDK, Claude Agent SDK, or another framework
It’s open-source (MIT license).
Link in the reply 👇
Working on an opensource keybound delegatable and revocable AI Authorization framework AGNAP. Agent Grant Negotiation and Authorization Protocol.
github.com/DaevMithran/agnap
#AISafety #Harness #GNAP
Daev Mithran retweeted
introducing @variancehouse, part of @theresidency.
we’re putting 14 of the most ambitious deep-tech builders in one house for 30 days.
$500k+ in credits, including $10k in anthropic credits for every resident.
with one of the craziest mentor lineup.
logan kilpatrick (gemini)
danielle strachman (1517 fund)
shyamal anadkat (ex-openai)
harshil mathur (razorpay)
harshita arora (yc partner)
terry winograd (stanford)
stay, food and workspace covered. no fees. no equity
& need based stipend
cohort starts from 15 sep–15 oct.
apply now, this one’s different
Attending @mcpsummit and @linuxfoundation Open Source Summit 2026 in Mumbai, India.
#ossummit #opensource #mcp
My second contribution to OpenClaw, now plugins can make LLM calls through OpenClaw's existing inference stack instead of calling provider APIs directly.
github.com/openclaw/openclaw…
Context Management using Hierarchical Clustering!
More details on: github.com/ekailabs/contexto…
OpenClaw VM a cross platform virtual machine for running OpenClaw on macOS and Linux. It wraps VirtualBox + Vagrant behind a friendly CLI tool. Ideal for hosting OpenClaw locally and development.
Repo: lnkd.in/gYW2u4jx
Feedback and contributions welcome!
#OpenClaw
Daev Mithran retweeted
Software Engineering is the only industry where 5 days of study is more important than 5 years of experience
Daev Mithran retweeted
Q4 Product Update: cheqd Studio📀
cheqd Studio now lets you fully manage credential & member lifecycles with Bitstring Status Lists: revoke, suspend, or customise statuses on-chain & off-chain.
Enterprise-ready SSI just got simpler.























