Usage Limits & Rationing

Demand-side rationing of AI capacity — provider-imposed limits and user habits for living within them — driven by compute scarcity and investor narratives.

Also known as: usage limits, rate limits, token rationing, rolling 5-hour window, peak hours, Agent SDK credit, throttling

How authors use it

Oldest first (extraction date when no publication date is known).

Documents (7)

• Fragments: July 6 Martin Fowler · 2026-07-06 source ↗ Reader 🖥️
  Agentic software development has crossed from aspiration into established practice — the debate is no longer whether it changes software engineering but how — and the central live question is whether classic design and architecture fundamentals still matter, with Fowler's emerging answer being that they do, both because agents use the same constructs humans do and because good design hedges against escalating token costs and the risk of losing affordable access to AI.
• Agentics / Tech Things: Tokenmaxxing is dead, long live tokenmaxxing theahura · 2026-06-27 Reader 🖥️
  Tokenmaxxing — deliberately spending large amounts of AI inference tokens — was born as a purposeful blunt-force tactic to force AI adoption past resistant employees and died once vendors withdrew subsidies and raised prices, but it will be reborn on genuine-ROI footing because a new 'compounding correctness' regime (more tokens now reliably yield better results, making unsupervised 24/7 agent loops worth running) plus cheap open models run in loops make maximal token spend economically rational rather than mere incentive-gaming.
• Developer Newsletter: higher usage limits, dreaming for Claude Managed Agents, and Claude Security Claude Team · 2026-05 Reader 🖥️
  Anthropic is repositioning Claude from a chat model into a managed agent platform: more compute headroom for Claude Code, persistent and self-improving agents ('dreaming'), enterprise-grade isolation (self-hosted sandboxes, MCP tunnels, AWS GA), and a security product (Claude Security beta) — together signaling that the next competitive surface is not the model but the governed runtime around it.
• Hard truths about building in the AI era | Keith Rabois (Khosla Ventures) Lenny Rachitsky, Keith Rabois · 2026-04-12 source ↗ Reader 🖥️
  Building a world-class company requires ruthless talent identification and density above all else, and in the AI era this means fewer but more capable people operating at extreme speed with business acumen replacing traditional role specialization.
• I stopped hitting Claude's usage limits - 10 things I changed kaize · 2026-04-10 Reader 🖥️
  You can avoid hitting Claude's usage limits by understanding that Claude counts tokens (not messages) and adopting ten habits that minimize unnecessary token consumption: editing instead of following up, starting fresh chats, batching questions, caching files in Projects, saving preferences to Memory, disabling unused features, choosing cheaper models for simple tasks, spreading work across the day, avoiding peak hours, and enabling overage billing.
• Starving Genies Kent Beck · 2026-04-03 source ↗ Reader 🖥️
  The simultaneous usage throttling by AI model providers is best understood not as an engineering constraint but as a narrative constraint: all three face the same investor-story shelf life, forcing synchronized demand-side rationing while the race to bend the supply curve up determines who wins the next wave.
• [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 stellaraccident (via https://github.com/anthropics/) · 2026-04-02 source ↗ Reader 🖥️
  Extended thinking tokens are structurally load-bearing for complex autonomous engineering workflows; Anthropic's staged reduction and redaction of thinking depth beginning in February 2026 caused a measurable, catastrophic quality regression that multiplied compute costs and destroyed a proven multi-agent workflow, and the report provides quantitative evidence to persuade Anthropic to restore deep thinking for power users.

Propositions touching this concept (8)

By publication month, oldest first (extraction month when no publication date is known).

2026-04

• implicitThe synchrony of the throttling is itself the strongest evidence for narrative as the constraint, because engineering constraints would produce variation across companies.
  Starving Genies · p4
• explicitUsage limits are investor-signaling devices: evidence of managing toward profitability, not admissions of resource exhaustion.
  Starving Genies · p5
• explicitThrottling may not be temporary — it could mark the beginning of Extract phase where heavy AI usage becomes a premium product.
  Starving Genies · p7
• explicitStarting March 26, 2026, Anthropic applies a higher rate-limit cost multiplier during weekday peak hours (5-11 AM PT).
  I stopped hitting Claude's usage limits - 10 things I changed · p10

2026-07

• explicitThe router runs headless in CI (GitHub Actions), which enables running tasks during off-peak hours to reduce API costs.
  GitHub - musistudio/claude-code-router: Use Claude Code as the foundation for coding infrastructure, allowing you to decide how to interact with the model while enjoying updates from Anthropic. · p7
• explicitBecause most STAR processing is I/O-bound (telemetry retrieval and LLM requests), an asynchronous, Redis-backed execution model lets it process multiple analyses concurrently while respecting external rate limits from Datadog and the internal AI gateway.
  Expedia Uses AI Driven Service Telemetry Analyzer to Accelerate Incident Investigation · p6

2026-09

• implicitThe parallel-execution latency argument implicitly assumes that hitting four or five stores/APIs simultaneously for every query is always operationally free — no rate limits, contention, or cascading load on the external APIs or the graph database under concurrent traffic.
  Enterprise Analytics Beyond Dashboards: Intelligent Data Orchestration with LLMs · p23
• implicitApparent redundancy — such as two APIs providing identical functionality — can actually be a coherent, deliberate design choice once organizational context is included (e.g. one API tuned for growing agentic traffic with different rate limits, the other for 'normal' traffic); a decision that looks incoherent from a pure infrastructure view can be coherent from an organizational view.
  Coherence, Connections, and… Spacetime Crystals · p12

Tensions

Open gaps (31)

• incompleteThe 'dreaming' mechanism is named and promised but not specified. We get no description of what is consolidated, what is discarded, how cross-user contamination is prevented, whether patterns are per-tenant or shared, whether 'dreams' are reviewable, or how a bad dream is unlearned. The cognitive metaphor is doing all of the persuasive work.
  Developer Newsletter: higher usage limits, dreaming for Claude Managed Agents, and Claude Security — read next: Anthropic's eventual technical post on the dreaming architecture; also: Voyager (Wang et al.) on skill libraries from agent experience; Reflexion (Shinn et al.) on self-reflection; recent literature on 'experience replay' for LLM agents; and any work on session-trace privacy for shared-tenant learning.
• incompleteClaude Security scans for vulnerabilities and patches them, but the newsletter is silent on how false positives are surfaced, how confidence ratings are calibrated, what training data the scanner uses, and what happens when the same model family produced the vulnerable code in the first place. The conflict-of-interest case is not addressed.
  Developer Newsletter: higher usage limits, dreaming for Claude Managed Agents, and Claude Security — read next: Compare against Snyk, Semgrep, GitHub Advanced Security, Endor Labs, and the recent wave of agentic SAST tools; read Anthropic's safety/eval team posts on adversarial code; also: literature on LLMs as code reviewers (CodeRabbit research, DeepMind's 'CompCoder' line).
• incomplete'Managed Agents' is the brand, but the newsletter never names the alternative it is positioning against: the open-source agent stack (LangGraph, AutoGen, OpenAI's Agents SDK, Microsoft's AutoGen, the MCP ecosystem itself). The trade-off between portability and a vendor-managed runtime is unexamined.
  Developer Newsletter: higher usage limits, dreaming for Claude Managed Agents, and Claude Security — read next: Compare with OpenAI's Agents SDK announcements, Google's Vertex AI Agent Builder, LangGraph Platform; recent essays on the agent runtime layer (Simon Willison, Swyx) and the 'agent OS' framing.
• incompleteThe pricing split — Agent SDK and `claude -p` no longer counting against Claude plan limits, but receiving a separate credit — is announced without numbers. This makes it impossible to evaluate whether the new metering is a substantive change or a reframing exercise.
  Developer Newsletter: higher usage limits, dreaming for Claude Managed Agents, and Claude Security — read next: Anthropic pricing page and the release notes linked from this newsletter; community write-ups after June 15 when the change takes effect.
• alternative explanation not consideredRegulatory signaling is not discussed. Providers may be synchronizing on usage limits partly to demonstrate responsible scaling to regulators, not only to investors.
  Starving Genies — read next: Look at EU AI Act timeline and responsible scaling commitments from Anthropic, Google, OpenAI.
• evidence gapThe 'narrative shelf-life' explanation is offered through elimination rather than direct evidence. No investor statements, board dynamics, or financial data are cited.
  Starving Genies — read next: Cross-reference with investor communications, earnings calls, or insider reporting on AI funding dynamics.
• incompleteCost and economics are absent. Multi-Claude parallelism, headless fan-out over thousands of files, and repeated 2-3x iteration all multiply token spend, yet the guide never mentions cost, usage/rate limits, or the threshold at which harness overhead pays for itself.
  Claude Code: Best practices for agentic coding — read next: Read against extract-anthropic-harness-design-long-running-application-development#p13 (the harness-vs-solo 20x cost multiplier and when it is justified), the ai-unit-economics and usage-limits-rationing concepts.
• alternative explanationsThe report attributes all quality regression to thinking depth reduction but does not consider alternative causes: model weight updates, RLHF tuning changes, system prompt modifications, or context window handling changes that may have occurred in the same period.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: Investigate Anthropic's model changelog for the February-March 2026 period to see what other changes coincided with the thinking reduction.
• confounding variablesThe March data conflates three simultaneous changes: (1) thinking reduction, (2) thinking redaction, and (3) the user's 5-10x scale-up of concurrent agents. The report acknowledges this but cannot cleanly separate the effects.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: A controlled experiment with fixed concurrency and variable thinking depth would isolate the causal factor.
• generalizabilityAll data comes from a single team's workflow (systems programming, MLIR, GPU drivers) with a specific CLAUDE.md configuration. It is unclear whether the regression equally affects other domains or simpler tasks.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: Cross-reference with other issue reports in the claude-code repository to see if the pattern is domain-specific or universal.
• missing counterfactualThe report does not test whether the same prompts produce better results with a different model (e.g., the 'other provider' mentioned). This would strengthen the claim that the issue is thinking-specific rather than prompt-specific.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: The report mentions switching to another provider that is 'doing superior quality work' but provides no comparative data.
• Anthropic's perspectiveThe report is entirely from the user's perspective. Anthropic's reasons for reducing/redacting thinking are not explored. There may be safety, privacy, or technical reasons that would reframe the tradeoffs.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: Look for Anthropic's public statements about thinking token policy, the redact-thinking header, and any response to this issue.
• Claude's self-analysis reliabilityThe report was produced by Claude analyzing its own logs, which raises questions about whether a model can objectively assess its own degradation, especially when the analysis itself may be affected by the same thinking constraints it is documenting.
  [MODEL] Claude Code is unusable for complex engineering tasks with the Feb updates #42796 — read next: Consider whether the analysis should be independently verified by humans or a different model to control for self-serving bias.
• uninformedThe thread assumes that Anthropic's session limit A/B testing is unusual or nefarious. In reality, nearly all SaaS products use feature flags and A/B testing for resource allocation. The finding is significant for transparency reasons but the conspiratorial framing ('tricking you') is unwarranted without evidence of deceptive intent.
  Claude Code Source Leak Megathread — read next: Standard practices in SaaS product management; Anthropic's Terms of Service on usage limits
• unsupported claimThe 50-70% budget savings from using Haiku is stated without evidence or calculation.
  I stopped hitting Claude's usage limits - 10 things I changed — read next: Look for Anthropic's published pricing ratios between Haiku, Sonnet, and Opus to verify.
• oversimplificationThe quadratic formula S*N(N+1)/2 assumes equal-length exchanges and ignores that Claude's responses vary greatly in length. It also conflates input tokens (which are re-read) with output tokens (which are not re-read).
  I stopped hitting Claude's usage limits - 10 things I changed — read next: Anthropic's documentation on how input vs output tokens are counted for rate limits.
• missing nuanceThe claim that Projects caching means files 'don't eat into your usage' is stated without caveats. Anthropic's actual caching behavior and whether cached tokens truly have zero rate-limit impact is not substantiated.
  I stopped hitting Claude's usage limits - 10 things I changed — read next: Anthropic's documentation on prompt caching and how cached tokens count toward limits.
• temporal sensitivitySeveral tips reference specific Anthropic platform features and policies (peak hours starting March 26, 2026; overage billing; model tiers) that may change. The advice is tightly coupled to the current product state.
  I stopped hitting Claude's usage limits - 10 things I changed — read next: Anthropic's changelog and pricing pages for current policies.
• missing considerationNo discussion of Claude Code, API usage, or non-web-UI contexts. All advice is implicitly scoped to the Claude web/desktop interface.
  I stopped hitting Claude's usage limits - 10 things I changed — read next: Separate guides for API and Claude Code token optimization.
• scopeNo discussion of pricing, rate limits, or scale constraints for the File Search tool
  Gemini API File Search is now multimodal: build efficient, verifiable RAG — read next: Gemini API pricing documentation
• incompleteNo discussion of cost model. Colab free-tier has usage limits and paid tiers have costs. For agent-driven workloads that may execute frequently and consume GPU time, the economics are entirely unexamined.
  Google Brings MCP Support to Colab, Enabling Cloud Execution for AI Agents — read next: Google Colab pricing documentation; analyses of compute costs for AI agent workflows
• incompleteThe claim that psychological safety is negatively correlated with success is stated as a blanket assertion without evidence beyond a sports analogy (Jordan Rules). No engagement with the substantial research literature on psychological safety in knowledge work, nor acknowledgment that survivorship bias may be at play (we see the winning teams that lacked it, not the many that failed).
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• incompleteThe anti-customer-feedback position is argued entirely through consumer examples (Porsche, Airbnb, DoorDash) but then extended to SMB without distinguishing where on the consumer-to-enterprise spectrum the advice flips. The Porsche example proves that stated preferences diverge from revealed preferences, not that all consumer feedback is harmful.
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• illogicalThe argument that public criticism optimizes for the system while private criticism optimizes for the individual assumes a binary choice. It does not address whether the information benefits of public criticism could be achieved through other means (e.g., transparent issue tracking, team retrospectives) without the potential costs of public shaming.
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• uninformedThe claim that AI writing quality degrades with length due to 'token rationing' by LLMs optimizing economics is presented as a confident theory but reflects a misunderstanding of how inference works. Output quality degradation with length has more to do with attention mechanisms, context window limitations, and the difficulty of maintaining coherence over long sequences than with deliberate economic throttling.
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• incompleteThe advice to skip senior hires and develop talent internally is supported by examples from companies that had extraordinary founding teams (Ramp, Trade Republic). The selection bias is unaddressed: these companies may have been able to develop internal talent precisely because their initial talent density was exceptional. The advice may not generalize to companies with weaker founding teams.
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• incompleteThe existential question about whether foundation model labs will absorb all value from application-layer AI startups is raised but never addressed. Given that Rabois is actively investing in AI companies, this is a significant gap — he identifies the question but offers no framework for how startups should think about durability against improving foundation models.
  Hard truths about building in the AI era | Keith Rabois (Khosla Ventures)
• incompleteCost is quantified, value is asserted. Every number in the piece is a cost (token prices, $12.5k/attempt, $600/month, $1000/day); the productivity or quality VALUE that would justify maximal spend is never measured, only hedged ('ostensibly'). The essay therefore cannot tell marginal from transformative benefit.
  Agentics / Tech Things: Tokenmaxxing is dead, long live tokenmaxxing — read next: Read HBR's 'How People Are Really Using AI in 2026' / the 'workslop'/'thinkslop' findings (marginal-not-game-changing benefits) and controlled productivity studies; ask whether the dev-loop value curve is transformative or merely marginal against unsubsidized cost.
• incompleteNo mechanism for the regime shift. The author asserts 'That's no longer true. We've entered a different regime' but never explains what technically flipped compounding error into compounding correctness — better base models? harness-level self-verification? evals in the loop? RL-from-verifiable-rewards? The rebirth's causal engine is a black box.
  Agentics / Tech Things: Tokenmaxxing is dead, long live tokenmaxxing — read next: Read on agent self-verification and eval-in-the-loop harness design, RLVR / verifier-based training, and what actually reduces error accumulation in long runs; distinguish model-capability gains from harness/scaffolding gains.
• incompleteVendor-vs-user value and the author's own incentive alignment are unexamined. 'More tokens = better' is exactly the narrative that benefits whoever sells tokens or AI-adoption services — and the author runs such a service, soliciting business throughout ('reach out, we can help!'). Who bears the cost (users, post-subsidy) versus who captures the value (vendors, consultants) is never separated, and the conflict is undisclosed.
  Agentics / Tech Things: Tokenmaxxing is dead, long live tokenmaxxing — read next: Read critically for supply-push vs demand-pull: who profits from maximal token consumption; compare against sources arguing 'most effective users are not the heaviest users'; watch for the token-seller framing in any pro-spend argument.
• incompleteThe two-kinds distinction is not integrated with the rebirth. The author says compounding correctness rescues developer loops but that pipeline agents are waste; he never reconciles whether compounding correctness would also rescue pipelines, or whether the generalist-platform shift moots custom pipelines regardless. The two 'this too shall pass' stories run in parallel without a unifying account.
  Agentics / Tech Things: Tokenmaxxing is dead, long live tokenmaxxing — read next: Trace whether reliability gains (compounding correctness) and consolidation gains (generalist platforms replacing bespoke agents) are the same phenomenon or two; read on why deterministic pipelines still beat agents for fixed high-accuracy tasks.