PodBrowser
Crypto Banter

NEAR Is Solving Crypto's BIGGEST Problem! | Illia Polosukhin

Saturday, 7 November 2026 · 3 min read · Listen to the episode ↗

NEAR co-founder Illia Polosukhin explains how NEAR Intents lets users request outcomes across Bitcoin, Ethereum, Solana, and other networks without choosing the executing chain. He discusses confidential transactions and why privacy could protect on-chain agents from exposing strategies or facing adversarial trading. The conversation also explores NEAR AI’s goal of private, verifiable inference and Polosukhin’s vision of user-owned AI and decentralized economic activity.

NEAR’s central proposition is that users and software agents should be able to specify what they want to do without managing the underlying blockchains. NEAR Intents coordinates transactions across networks including Bitcoin, Ethereum, and Solana, letting users request outcomes such as acquiring Bitcoin or earning yield without choosing the chain that executes them. Illia Polosukhin said the product launched in November 2024 and that product-market fit became apparent around November 2025 as Zcash use grew. The project began with a different but related problem: NEAR started as an AI company, and its blockchain emerged as a payments layer for compensating people who supplied training data.

Intents also supports confidential asset movement. A user can exchange Bitcoin for shielded Zcash without a NEAR account, or move shielded Zcash into Bitcoin or USDT on Tron to spend without publicly exposing a balance. Early integrations included Infinax, Swapkit, Trust Wallet, and Ledger. The near.com interface brings trading, earning, payments, and confidential transactions across assets and chains together, and can withdraw from an earn vault to make a payment. Polosukhin said NEAR has extended confidentiality to USDT, USDC, Bitcoin, Zcash, Solana, and Ethereum, and to activities including trading, payments, stocks, and perpetuals.

Privacy remains incomplete in some applications. On Hyperliquid, NEAR currently hides the origin and owner of a position, but not the position itself; concealing the position is a future goal. Polosukhin also sees privacy as a defense for on-chain agents, which might otherwise expose trading information or face MEV, sandwich attacks, and other adversarial trading. He expects agents on personal devices, as well as in businesses, factories, and vehicles, to transact and coordinate economic activity. Blockchain could help them find counterparties and settle activity, while NEAR Intents could route transactions under wallet policies that restrict spending, addresses, and asset types.

Polosukhin expects agents to use multiple currencies. Stablecoins are useful because dollars are the unit of account for costs such as GPUs, while staking NEAR can provide access to inference. He wants NEAR to become an AI asset that backs compute, but said most settlement is currently in dollars and NEAR may be more useful where compute or trust is needed. He identified a frontier AI lab becoming a singular centralized power as the greatest threat to freedom and decentralized economic systems: a company could use model responses to influence users at scale. He considers that more likely than literal Doomsday, though he views the singularity scenario as low probability because people want plurality.

NEAR AI is intended to offer private, verifiable inference, so inference and GPU providers cannot see agent inputs or alter outputs. Model and serving details are verified, hashed, and hardware-signed. Polosukhin said Venice can use NEAR AI inference when users select its encrypted option; Venice’s alternative zero-data-retention policy alone does not guarantee privacy. He sees open-source and open-weight AI, alongside user-owned, private, verifiable systems, as ways to decentralize access. NEAR began as NEAR AI after Polosukhin left Google in 2017; at Google he had implemented an early transformer approach for natural-language understanding, and he expected major AI advances to arrive sooner than they did.

On quantum computing, Polosukhin said practical threats remain years away and would require substantial hardware, capital, and research, likely beginning in a few labs. He cited a system with about 500 qubits but said recovering a private key would still take time, identifying Bitcoin as the main cryptocurrency risk. NEAR was designed in 2018 so its cryptography could be changed, and its account model allows key rotation without moving assets. It offers post-quantum signing with ML-DSA, including an option requiring both elliptic-curve and ML-DSA signatures; Meteor Wallet supports one-click rotation. He expects Bitcoin to add post-quantum signatures, calling it mainly a coordination issue, but said unupgraded Bitcoin could become inaccessible and Satoshi’s coins could be bricked. NEAR’s mission remains giving people ownership of assets, data, and choices; Polosukhin said Intents has product-market fit and NEAR is “getting there” on AI, with a goal of putting the full stack in the hands of billions.

This summary was generated from the episode transcript and can contain mistakes.