cuppa

brewed 03:39 AM
← May 13

Thursday

may14

2026

May 15 →

the brief

Anthropic redrew the line between interactive and programmatic use: dedicated monthly SDK credits start June 15 even as Claude Code limits get a short-term bump. OpenAI kept pressing Codex into the enterprise with a Windows sandbox deep-dive and a switcher offer. Tooling saw steady updates, and security-focused evals moved fast, with new leaderboard shifts and range completions.

the poursit · sip · 12 items

alerts

(01)
  • hn/frontpage· feedMay 13, 09:40 PM

    Reports of Claude Design data loss

    HN users report losing access to past Claude Design projects after unsubscribing; export or back up artifacts before changing plans until policies are clarified.

    Tell HN: Dont use Claude Design, lost access to my projects after unsubscribing — I wanted to try codex after 5 months of claude code max subscription. And then I went back to my previous projects on claude design only to realize I don't have access to them anymore. This is a first. I never lost access to any of my past sessions because I unsubscribed in any of the LLM apps. I actually wanted to try out codex previously, but had similar experience with my credits. They gave extra credits equi...

    signal 6hype 2bug_reportclaudesubscriptionsource ↗

pulse

(07)
  • @unknownMay 13, 07:07 PM

    Claude Code weekly limits up 50%

    Anthropic temporarily raised Claude Code weekly limits through July 13, stacking with the recent 2x 5‑hour boost, giving builders more headroom for AFK coding work.

    Claude Code weekly limits are increasing 50%, now through July 13. Live now for all Pro, Max, Team, and seat-based Enterprise users. pic.x.com/5nU0XX4RZY

    signal 9hype 1product_updatequota_increaseclaude_codesource ↗
  • @unknownMay 13, 05:10 PM

    Claude adds monthly SDK credits

    Starting June 15, paid Claude plans get a separate monthly budget for programmatic use across the Agent SDK, claude -p, GitHub Actions, and third‑party tools, decoupling it from chat limits.

    Starting June 15, paid Claude plans can claim a dedicated monthly credit for programmatic usage. The credit covers usage of: - Claude Agent SDK - claude -p - Claude Code GitHub Actions - Third-party apps built on the Agent SDK

    signal 8hype 1pricing_updateusage_creditsagent_sdksource ↗
  • @OpenAIDevsMay 14, 01:26 AM

    Inside Codex Windows sandbox

    OpenAI details a Windows sandbox that lets coding agents stay useful without full machine access, outlining isolation, permissions, and UX tradeoffs for safe agent execution on Windows.

    To bring Codex to Windows, we had to answer a hard question: how do you let coding agents stay useful without forcing developers to choose between constant approval prompts and full machine access? Here’s how we built the Windows sandbox for Codex: openai.com/index/building…

    signal 8hype 1sandboxingagent_securitywindowssource ↗
  • @unknownMay 13, 03:35 PM

    Codex offers 2 months for switchers

    OpenAI is courting enterprises to adopt Codex with an official pathway and two free months for new users, aiming to formalize agent use at work.

    Want to (officially) use Codex at work? Send this post to your CTO to bring your team to Codex. Eligible enterprise customers who switch in the next 30 days get 2 free months of Codex usage for new users. pic.x.com/38e8y7MAmg

  • @steipeteMay 13, 10:30 PM

    Crabbox 0.13 ships sandbox runs

    Crabbox adds Modal sandbox runs, native Windows script and preflight support, full resync for stale workdirs, and clearer SSH/sync failure hints to streamline PR workflows.

    Crabbox 0.13.0 is live 🧪 Modal sandbox runs 🧼 Full resync for stale workdirs 🪟 Native Windows script + preflight support 🔧 Clearer SSH/sync failure hints Been using that for almost every PR now. github.com/openclaw/crabb…

    signal 6hype 2version_releasedev_toolingrelease_notessource ↗
  • @vercel_devMay 13, 05:29 PM

    Vercel Gateway adds Opus fast

    Vercel’s AI Gateway now supports Opus 4.7 fast mode via a simple speed: 'fast' API flag, offering lower latency without manual routing changes.

    Fast mode for Opus 4.7 is live on AI Gateway. Try it via API: 𝚜𝚙𝚎𝚎𝚍: '𝚏𝚊𝚜𝚝' vercel.com/changelog/fast…

    signal 6hype 1ai_gatewayfeature_updateapi_paramsource ↗
  • anthropics/claude-code· feedMay 13, 11:19 PM

    Claude Code v2.1.141 released

    Release adds terminalSequence in hook output, HTTPS-preferred plugin cloning, workspace-scoped federation via ANTHROPIC_WORKSPACE_ID, and env flags for smoother SSH-free setups.

    v2.1.141 — What's changed Added terminalSequence field to hook JSON output so hooks can emit desktop notifications, window titles, and bells without a controlling terminal Added CLAUDE_CODE_PLUGIN_PREFER_HTTPS to clone GitHub plugin sources over HTTPS instead of SSH, for environments without a GitHub SSH key Added ANTHROPIC_WORKSPACE_ID environment variable for workload identity federation — scopes the minted token to a specific workspace when the federation rule covers more than one Added cl...

    signal 9hype 0release_notesclaude_codehookssource ↗

findings

(02)
  • @unknownMay 13, 09:59 PM

    MDASH tops vuln discovery leaderboard

    An updated leaderboard shows MDASH at #1 using a multi-model approach, signaling rapid advances in AI-driven vulnerability discovery and defense.

    cybergym.io just updated its leaderboard and MDASH is now #1 using a new multi-model approach. Huge credit to Taesoo Kim and the Autonomous Code Security team for pushing the frontier on AI-driven vulnerability discovery and defense microsoft.com/en-us/security… pic.x.com/m6zlkaVYwJ

    signal 6hype 2leaderboardsecurityvulnerability_discoverysource ↗
  • @bchernyMay 13, 05:40 PM

    Mythos solves AISI cyber ranges

    UK AISI reports Mythos Preview as the first model to complete both its cyber ranges, including the previously unsolved Cooling Tower, pointing to meaningful gains for defenders.

    The UK AISI found Mythos Preview is the first model to solve both their cyber ranges end-to-end. No model had ever solved the AISI’s “Cooling Tower” cyber range before. We're getting it to defenders as fast as we responsibly can. More to come on our Glasswing work soon. x.com/logangraham/st…

    signal 8hype 2model_evaluationbenchmarkcybersecuritysource ↗

voices

(02)
  • @mattpocockukMay 13, 08:09 PM

    Programmatic cuts behind monthly credit

    Matt Pocock argues the new dedicated credit pool effectively reduces AFK/programmatic limits by roughly 5–20x for many, urging teams to reassess workflows.

    Anthropic has given us a "dedicated monthly credit" Which, in effect, slashes AFK usage limits of Claude Code by ~5-20X Here's how it affects you: pic.x.com/1nCCheqEC4

    signal 8hype 1anthropicclaude_codeusage_limitssource ↗
  • @unknownMay 13, 06:57 PM

    A 10x cut to claude -p

    A contrarian take calls the change a disguised 10x reduction for claude -p usage, suggesting Anthropic is steering away from programmatic use and nudging users toward alternatives like Codex.

    This is the clarity we've been crying out for. But it's a poisoned chalice. This is a 10X cut to claude -p disguised as a monthly bonus. Anthropic is discouraging any kind of programmatic usage. And that's fine - no subsidy lasts forever. But it's time to try Codex. x.com/ClaudeDevs/sta…

    signal 5hype 4anthropicpricing_changeapi_limitssource ↗