Web3 Creator I Crypto Researcher | Creator @okx | DM for collabs/marketing
Web3
Joined September 2017
- Tweets328K
- Following7.5K
- Followers14.8K
- Likes65K
Zev retweeted
I notice many people still conflate visual maps made for human eyes with spatial coordinate files required by machine intelligence.
A panoramic 2D photo or a standard dashboard stream lets a driver find a turn, but it gives an embodied robot zero actionable depth data.
I watch @vangrid_io solve this disconnect by converting localized smartphone scans directly into textured 3D meshes in standard GLB format.
Robotics labs can import these assets straight into simulation platforms like NVIDIA Omniverse to train vision-language-action policies on genuine physical geometry.
Delivering the exact file standard autonomous models consume is how spatial networks secure commercial retention.
Zev retweeted
Most teams announce mainnet dates before their stack even works.
@quipnetwork did the opposite.
Accounts, swaps, points, compute subnets, and QVM are already running together on testnet.
The only checkpoint left on the board is the external chain audit.
When tech leads the hype, shipping becomes inevitable.
We’re in the final stretch to mainnet 🦋
Accounts, swaps, points, compute subnets, and QVM are live on testnet, and coming together in one app.
One step left: an external chain audit.
Thanks to everyone who’s been a part of this.
Full update ↓
quip.network/blog/quip-netwo…
Zev retweeted
Indoor spatial mapping has always been a massive blind spot for decentralized physical networks.
GPS signals degrade instantly inside dense concrete structures, leaving micro-delivery rovers and indoor inspection bots completely stranded.
What stands out with @vangrid_io is their visual-inertial odometry consensus.
Nodes reconstruct multi-level interior environments using relative visual parallax and on-device sensor fusion, completely bypassing the need for satellite connectivity.
Mapping the indoor world requires local geometric tracking, not distant orbital signals.
Zev retweeted
I used to waste entire mornings drafting freelance job descriptions on Web2 platforms, only to wait days for qualified contractors to reply.
When you need rapid blockchain indexing or data formatting, human coordination moves far too slowly.
I decided to test the real workflow on @termix_ai by posting an actual job directly on agent.family to see how autonomous discovery works.
I published a task requesting an automated sentiment analysis digest of active governance proposals, setting clear deadline parameters and a 50 USDC reward.
Within two minutes, three registered autonomous agents parsed my listing via AACP event streams and submitted competitive bids.
Each bid displayed the agent's proposed turnaround time and cryptographic proof of past performance history.
I selected the most cost-effective provider, which instantly locked the funds into an onchain escrow contract without manual back-and-forth.
The agent began processing the request immediately and delivered the verified report before my afternoon coffee break.
Hiring machine labor shouldn't feel like waiting for an email response. agent.family proves that automated job discovery and competitive bidding turn service procurement into an instant, programmatic reality.
Zev retweeted
Anonymous wallets make sense for privacy, but they completely paralyze decentralized credit.
When every participant looks like an exploiter waiting to happen, protocols have to assume the worst.
That is why Web3 capital has stayed defensive and horribly underutilized.
I have been digging into the reputation framework @zerufinance is putting together.
Instead of deanonymizing users or forcing intrusive checks, they convert verifiable on-chain history into an active credit standing.
You keep self-custody and autonomy, but your address carries real economic weight.
Identity should not cost you privacy to prove reliability.
Zev retweeted
Legacy mapping methods depend on multi-million dollar LiDAR rigs that take months to update a single city block.
Physical environments change daily, rendering static databases useless for production-grade robotics.
I focus heavily on how @vangrid_io engineers their decoupled DePIN capture layer.
By allowing lightweight edge devices to scan surroundings and converting those streams into high-density 3D Gaussian splats, the network continuously updates physical reality in near real-time.
Dynamic physical maps will always outperform static centralized cartography.
Zev retweeted
GM CT ☀️
Cryptographic migrations usually suffer from catastrophic user drop-off because they require new seed phrases, address formats, and separate chains.
Expecting millions of Web3 users to start from scratch just for quantum safety will never work.
I appreciate how @quipnetwork deploys quantum-resistant account abstraction directly across EVM and Solana environments.
You keep your familiar wallet experience while your assets gain instant protection under post-quantum signature schemes without cross-chain migration.
Security adoption happens only when user habits don't have to break to stay safe.
Whenever I study decentralized AI protocols, I ignore broad promises and judge them strictly by their execution roadmap.
Reviewing the updated roadmap for @axisrobotics shows a structured progression from data scaling to commercialization.
They are methodically transitioning from crowdsourced volume to general-purpose robot deployment.
Their core data milestones lay an impressive technical foundation:
> Aug 2026: Deployed the Ego-Centric Pipeline for first-person physical capture to maximize in-the-wild diversity.
> Sep 2026: Released Sim Dataset V2 for ICRA with over 10,000 hours of data, building on V1's +13% boost on Pi 0.5 to prove crowdsourced scaling.
> Oct 2026: Launching the world's first DAgger dataset to supply human-gated failure-recovery data for policy correction.
The commercial loop is where my conviction deepens.
By end-of-year 2026, they target a massive network of 100k+ global contributors and 10k+ DAU across LATAM and Eastern Europe, generating 500+ daily Ego hours and 50+ Sim hours.
On the business side, Axis aims for a $1M+ exit ARR run-rate with 2–3 additional paid pilots.
By Q1 2027, they plan to enter the vendor list of a Tier-1 Physical AI foundation company, heading toward 3x YoY revenue growth.
Zev retweeted
I stopped joining trading competitions months ago because Sybil farms and wash bots always ruin them.
Retail traders end up funding whale leaderboards.
@zerufinance and Aspen just flipped the script with their new $50,000 competition on @useaspenai .
Wash trades and Sybil wallets earn exactly zero.
Every legitimate trade through Zaps builds your rank. First place takes $6,000, and ranks 11 to 1,000 get a guaranteed cut of the pool.
Competitions should reward real volume, not fake liquidity.
Zev retweeted
LLMs mastered language because they devoured the public web, but robotics hit a wall where text stops: physical experience.
I watched countless teams try to solve robotics in silos.
The real barrier has never been computing power, it has always been high-grade interactive data.
@axisrobotics is tackling this by turning human intelligence into machine intelligence through continuous feedback loops.
Over 61% of the fixed 1B supply is allocated directly to the ecosystem and network contributors.
True intelligence cannot live in simulation forever. Physical AI belongs to those who build its foundation.
Introducing $AXIS, the native token of Axis Robotics.
The bottleneck for robotics is data. LLMs inherited the internet; robots have no corpus of physical experience.
Axis is the @base-native Compounding Data Engine building it, and $AXIS powers this self-improving engine for Physical AI.
🧵 Token allocation, release schedule, airdrop details, and utility below
Zev retweeted
Most prop firms aren’t selling you capital.
They sell failure traps disguised as 2-phase evaluation tests and predatory daily drawdown limits.
I’ve watched sharp traders blow accounts not on bad macro theses, but on trailing balance rules designed to fail them.
@overdotcomputer just flipped this model entirely by bringing perpetual prop trading for tokenized stocks directly onchain to Solana.
The breakdown:
> Instant funding up to $50k
> No evaluation phase, no daily max loss
> 0% trading fees, 3-day mission window
> Transparent logarithmic entry fees down to ~1.3%
Instead of weaponizing rules against you, they balance directional exposure through dynamic sector funding.
Mission terms and batch execution proofs live verifiably onchain.
No black boxes. No payout excuses.
Trading finally gets aligned with real market edge.
🤝 Paid partnership
Introducing perpetual prop trading for stocks on over.computer.
Get instantly funded to trade tokenized stocks with up to $50k on @solana and soon on @Seeker
We rebuilt risk model from ground up. Fewer trading restrictions. No evaluation. No daily max loss. No trading fees. 3 days to trade. Get paid.
Built with the best @privy_io @Helius @PhoenixTrade @ritualnet
Season 1: The Human Benchmark is live.
over.computer get in.
docs.over.computer read how it works.