关于生成Token的成本与效率，很多人心中都有不少疑问。本文将从专业角度出发，逐一为您解答最核心的问题。
问：关于生成Token的成本与效率的核心要素，专家怎么看？ 答：You are getting early access to this article as a subscriber. Your support makes articles like this possible. Thank you.I surveyed 112 major source-available projects to understand their AI contribution policy and whether or not they have actually accepted explicitly-labeled AI contributions. This survey included programming language implementations, databases, web browsers, programming libraries, operating systems, applications, and infrastructure projects.Of the 112 projects surveyed, only 4 banned AI contributions entirely: Zig, NetBSD, GIMP, and qemu. Some, like DuckDB and Elasticsearch, have commits by AI tools but have policies that seem to ban AI-assisted contributions. Perhaps only core contributors are trusted.Of the 112 projects surveyed, 70 had commits that explicitly mentioned being written with an AI assistant such as Claude, Gemini, Cursor, or Codex. Many more beyond those 70 projects had contribution policies that explicitly allowed AI contributions, but commits that were AI-assisted were not obvious.Across the board, low-level and high-level projects have adopted AI-assisted programming. Low-level and high-level projects have banned AI. There is not much of a pattern to be found.This survey is a good-faith attempt to determine policies and practices from public information. "Not found" in the data below basically means "no", but it's possible I just missed it. Please message me in that case.Furthermore, this survey makes no judgement at all about the use or disuse of AI, makes no judgement about having an explicit AI contribution policy, and makes no judgement about a preferred AI provider.Language ImplementationsProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerBunYesYesNot foundClaude, CursorC# (Roslyn)Not foundNoNot foundNoCPythonYesYesYesNoGCCNot foundNoNoNoGoYesYesNoNoHaskellYesYesNot foundNoKotlinYesYesNot foundClaudeLLVMYesYesYesNoNodeJSYesNot itself, but yes via V8Not foundClaude, GeminiOCamlYesYesYesNoPerlNot foundNoNot foundNoPHPNot foundNoNot foundNoRNot foundNoNot foundNoRubyYesYesNot foundNoTypeScriptNot foundNoNot foundNoZigNoNoYesNoWeb BrowsersProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerChromiumYesYesNot foundClaude, GeminiFirefoxYesYesYesClaude, Codex, ZedLadybirdYesYesYesNoWebkit (Safari)YesYesNot foundCodex, GeminiDatabasesProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerCassandraNot foundNoNot foundNoChromaYesYesNot foundNoClickHouseYesYesYesClaudeCockroachDBYesYesNot foundClaudeDuckDBOnly by contributorsYesYesNoElasticsearchYesYesYes (they appear not to allow AI contributions, but they have accepted them already)NoetcdYesYesNot foundNoHadoopYesYesDisclosure is part of pull request templateNoInfluxDBYesYesNot foundClaudeKafkaYesYesNot foundNoKVRocksYesYesYesNoMariaDBYesYes (specifically)Not foundNoMeilisearchYesNoYesNoMilvusYesYesNot foundClaudeMongoDBYesYesNot foundClaudeMySQLNot foundNoNot foundNoOpensearchYesNoMight be comingNoPostgreSQLNot foundNoNot foundNoPrometheusYesYesNot foundNoQdrantYesYesYesNoRedpandaYesYesNot foundClaudeRedisYesYesNot foundNoRocksDBYesYesNot foundClaudeScyllaDBYesYesNot foundClaude (via Seastar)SparkYesYesNot foundNoSQLiteNot foundNoNot foundNoTantivyYesYesNot foundClaudeTursoYesYesNot foundClaude, CodexWeaviateYesYesNot foundClaudeOperating SystemsProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerFreeBSDNot foundNoMight be comingNoLinuxYesYesMight be comingNoNetBSDNoNoYesNoXNU (Darwin)Not foundNoNot foundNoLibrariesProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerASP.NETNot foundNoNot foundNocurlYesNoYesNoDataFusionYesYesYesNoDjangoYesYesYesNoFastAPIYesYesYesNoFFmpegNot foundNoNot foundNogRPCNot foundNoNot foundGeminiJAXYesYesYesNollama.cppYes (AI-assisted)YesYesClaude, GeminiliburingYesYesNot foundNokerasYesYesNot foundGeminiNettyNot foundNoNot foundNonext.jsYesYesNot foundClaude, CursoropenSSLYesYesMight be comingNoPandasYesYesYesNoPolarsYesYesYesNoprismaYesYesNot foundClaude, GeminipydanticNot foundNoNot official but seemingly againstNoPytorchYesYesYesClaudeReactYesYesNot foundClaudeSpring BootNot foundNoNot foundNosvelteYesYesNot foundNotokioYesYesNot foundNothree.jsYesYesNot foundNouvYesYesNot foundClaudevLLMYesYesNot foundGeminivue.jsNot foundNoNot foundNoApplicationsProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerAudacityNot foundNoNot foundNoBlenderYes, for GSoC at leastNoMight be comingNoDarktableNot foundNoNot foundNoDrupalYesNoYesNoGiteaYesYesYesClaudeGitLabYesYesYesNoGIMPNoNoYes, for students at leastNoLibreOfficeNot foundNoNot foundNoMatomoNot foundNoNot foundNoMattermostYesYesNot foundClaudeOBS StudioNot foundNoYesNopi-holeNot foundNoNot foundNoThunderbirdNot foundNoNot foundNoVSCodeYesYesNot foundClaudeWordpressNot foundNoYesClaudeZulipYesYesYesClaudeInfrastructureProjectAllows AI contributionsHas accepted explicitly-labeled AI contributionsHas explicit AI contribution policyHas preferred AI providerCaddyYesYesYesNoCephYesYesNot foundNociliumYesYesNot foundNoenvoyYesYesYesNohaproxyNot foundNoNot foundNoIstioYesYesMight be comingNojaegerYesYesNot foundClaudeKeycloakYesYesNot foundNoKubernetesYesYesYesNomoby (Docker)YesYesNot foundNonginxNot foundNoNot foundNoOpen Policy AgentNot foundNoNot foundClaudeopenzfsYesYesNot foundNoqemuNoNoYesNosystemdYesYesYesNotraefikNot foundNoNot foundNoVaultNot foundNoNot foundNo
，这一点在WhatsApp网页版中也有详细论述
问：当前生成Token的成本与效率面临的主要挑战是什么？ 答：Complexity is the strategy. If the rules sound uncertain enough, people keep asking. And if people keep asking, licensing fees keep flowing, even after the copyrights have expired. Confusion, after all, is far cheaper than litigation.
来自产业链上下游的反馈一致表明，市场需求端正释放出强劲的增长信号，供给侧改革成效初显。
。TikTok粉丝,海外抖音粉丝,短视频涨粉是该领域的重要参考
问：生成Token的成本与效率未来的发展方向如何？ 答：试想：若小红书所有头部创作者均为平台雇员，其商业模式能否持续？
问：普通人应该如何看待生成Token的成本与效率的变化？ 答：杰夫：无暇关注淘汰者。更需聚焦时代变革节奏，毕竟下一个可能被淘汰的就是我们自身。危机感促使我们向前看。，这一点在金山文档中也有详细论述
展望未来，生成Token的成本与效率的发展趋势值得持续关注。专家建议，各方应加强协作创新，共同推动行业向更加健康、可持续的方向发展。