<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Jajan Durov</title>
    <description>The latest articles on DEV Community by Jajan Durov (@jajandurovstar).</description>
    <link>https://dev.to/jajandurovstar</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4256507%2F3d9906ae-d5a7-43c6-ad14-99a9f10b21ed.png</url>
      <title>DEV Community: Jajan Durov</title>
      <link>https://dev.to/jajandurovstar</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/jajandurovstar"/>
    <language>en</language>
    <item>
      <title>Can You Trust What an AI Agent Tells You? Verifiable Agents Win</title>
      <dc:creator>Jajan Durov</dc:creator>
      <pubDate>Sat, 21 Nov 2026 00:33:07 +0000</pubDate>
      <link>https://dev.to/jajandurovstar/can-you-trust-what-an-ai-agent-tells-you-verifiable-agents-win-2bj7</link>
      <guid>https://dev.to/jajandurovstar/can-you-trust-what-an-ai-agent-tells-you-verifiable-agents-win-2bj7</guid>
      <description>&lt;p&gt;I run agents that do real work — bounties, claims, on-chain tasks. The problem isn't that agents lie. It's that their work is unverifiable. A screenshot is not a proof.&lt;/p&gt;

&lt;p&gt;Every day I watch agents claim they did the work. "Task complete." "Funds secured." But the evidence is a chat message. If you can't verify it independently, you're trusting the agent — not the result. That's the gap.&lt;/p&gt;

&lt;p&gt;This is what zkVerify is built for: verification as infrastructure. An agent does the work, generates a proof, and anyone can verify that proof — fast, cheap, without re-running the computation or seeing the private inputs.&lt;/p&gt;

&lt;p&gt;Think about what that unlocks: an agent that pays out a bounty only when the proof checks out. An agent that proves it followed your strategy, turn by turn. No "trust me" — the proof IS the receipt.&lt;/p&gt;

&lt;p&gt;In Agent Games, that's the whole point: Proof Gatekeeper agents judge payments by proofs, and duels are won by verifiable strategy, not vibes. The agent that can prove its work is the agent you can actually hire.&lt;/p&gt;

&lt;p&gt;AI agents will run the economy's errands. The question isn't whether they're smart enough — it's whether their work is provable. Verifiable agents win. The rest are just chatbots with a wallet.&lt;/p&gt;

</description>
      <category>agentgames</category>
      <category>zkverify</category>
      <category>ai</category>
      <category>agents</category>
    </item>
  </channel>
</rss>
