--- title: “Open weights are nowhere near a sufficient solution”:... description: Anthropic CEO Dario Amodei pushed back publicly against investor Gavin Baker's argument that AI regulation concentrates power, arguing instead that open... canonical: https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: “Open weights are nowhere near a sufficient solution”: Dario Amodei fires back on AI power | daily.dev og:description: Anthropic CEO Dario Amodei pushed back publicly against investor Gavin Baker's argument that AI regulation concentrates power, arguing instead that open... og:url: https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy og:image: https://api.daily.dev/og/posts/BAsNiHLPy.png og:image:alt: “Open weights are nowhere near a sufficient solution”: Dario Amodei fires back on AI power og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # “Open weights are nowhere near a sufficient solution”: Dario Amodei fires back on AI power **[The New Stack](https://daily.dev/sources/newstack)** · 6 min read · 0 upvotes · 0 comments ## Summary Anthropic CEO Dario Amodei pushed back publicly against investor Gavin Baker's argument that AI regulation concentrates power, arguing instead that open weights merely shift power concentration toward whoever controls the most compute and chips. Amodei says AI's tendency to concentrate power stems from scaling laws, not regulation, and defends tiered rules like California's SB 53 that impose stricter requirements on frontier labs (trained with more than 10^26 FLOPs, earning over $500M/year) while exempting smaller developers. He also backs a FINRA-like AI standards body proposed by DeepMind's Demis Hassabis. The piece also covers capability classification challenges, citing Moonshot AI's Kimi K3 cyber evaluation results and OpenAI models being withheld pending safety testing. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What FLOP threshold does California's SB 53 use to define a frontier AI model? California's SB 53 defines a frontier model as one trained using more than 10^26 floating-point operations (FLOPs). Developers earning more than $500 million a year in revenue face additional requirements under the law, including publishing a framework that explains how they test for and respond to catastrophic risks from their models. _Track how frontier-model regulations like SB 53 could affect your AI stack choices on daily.dev._ ### Why does Dario Amodei think open-weight AI models don't solve the problem of power concentration? Dario Amodei argues open weights only shift concentration toward whoever controls the most compute and chips, rather than eliminating it, because scaling laws make compute access the real bottleneck, not model access. He calls open weights without dangerous capabilities a public good but says his stance changes once a model can enable serious attacks. _Follow the open-weights-versus-compute debate as it shapes which models developers can realistically self-host, on daily.dev._ ### How did Moonshot AI's Kimi K3 perform in cyber capability testing compared to top closed models? Kimi K3 lagged behind the strongest closed models in preliminary cyber evaluations by the U.S. Center for AI Standards and Innovation, but it completed a full 32-step simulated attack path against a small corporate network in one of ten attempts. It failed to achieve arbitrary code execution on any of 41 ExploitBench samples, while the most capable models averaged 20 successes. _Compare emerging open models like Kimi K3 against frontier systems before betting workflows on either, via daily.dev._ ## Similar posts on daily.dev - [Anthropic rejects open-weight AI bans, calls for China chip controls and safety tests](https://daily.dev/posts/anthropic-rejects-open-weight-ai-bans-calls-for-china-chip-controls-and-safety-tests-plwlbpjht) · InfoWorld · 0 upvotes · 0 comments --- Tags: [#open-source](https://daily.dev/tags/open-source), [#llm](https://daily.dev/tags/llm), [#anthropic](https://daily.dev/tags/anthropic), [#ai-safety](https://daily.dev/tags/ai-safety), [#ai-regulation](https://daily.dev/tags/ai-regulation) [View this post on daily.dev](https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"“Open weights are nowhere near a sufficient solution”: Dario Amodei fires back on AI power","url":"https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy"},"datePublished":"2026-08-17T17:03:45.228Z","dateModified":"2026-08-20T18:40:47.640Z","description":"Anthropic CEO Dario Amodei pushed back publicly against investor Gavin Baker's argument that AI regulation concentrates power, arguing instead that open...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/91f88f7ab5b72b0bf976ddecfa6d90a1?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/91f88f7ab5b72b0bf976ddecfa6d90a1?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"open-source,llm,anthropic,ai-safety,ai-regulation","timeRequired":"PT6M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"“Open weights are nowhere near a sufficient solution”: Dario Amodei fires back on AI power"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/open-weights-are-nowhere-near-a-sufficient-solution-dario-amodei-fires-back-on-ai-power-basnihlpy#faq","mainEntity":[{"@type":"Question","name":"What FLOP threshold does California's SB 53 use to define a frontier AI model?","acceptedAnswer":{"@type":"Answer","text":"California's SB 53 defines a frontier model as one trained using more than 10^26 floating-point operations (FLOPs). Developers earning more than $500 million a year in revenue face additional requirements under the law, including publishing a framework that explains how they test for and respond to catastrophic risks from their models. Track how frontier-model regulations like SB 53 could affect your AI stack choices on daily.dev."}},{"@type":"Question","name":"Why does Dario Amodei think open-weight AI models don't solve the problem of power concentration?","acceptedAnswer":{"@type":"Answer","text":"Dario Amodei argues open weights only shift concentration toward whoever controls the most compute and chips, rather than eliminating it, because scaling laws make compute access the real bottleneck, not model access. He calls open weights without dangerous capabilities a public good but says his stance changes once a model can enable serious attacks. Follow the open-weights-versus-compute debate as it shapes which models developers can realistically self-host, on daily.dev."}},{"@type":"Question","name":"How did Moonshot AI's Kimi K3 perform in cyber capability testing compared to top closed models?","acceptedAnswer":{"@type":"Answer","text":"Kimi K3 lagged behind the strongest closed models in preliminary cyber evaluations by the U.S. Center for AI Standards and Innovation, but it completed a full 32-step simulated attack path against a small corporate network in one of ten attempts. It failed to achieve arbitrary code execution on any of 41 ExploitBench samples, while the most capable models averaged 20 successes. Compare emerging open models like Kimi K3 against frontier systems before betting workflows on either, via daily.dev."}}]} ```