Anthropic SDKs add a message threads beta the API docs don't list yet

AnthropicWhat's new#

Anthropic's Python and TypeScript SDKs shipped client.beta.messages.thread() on Oct 23 with no announcement. The helper keeps the conversation on the server so each request sends only new messages, but the API docs still call the Messages API stateless, and Anthropic hasn't said which accounts or models accept it. (confirmed for what the SDKs ship)

AnthropicHow people are using it#

Weekend Hark posts name two more copyable jobs: filming the pantry for a weekly priced restock, and recording conference livestreams for session notes, plus usage and anti-bot walls to plan around. User reports; we have not tested Hark. Saturday's Ferriss thread roundup has the earlier jobs.

AnthropicThe digest#

Anthropic#

• Users report Agent SDK and claude -p burn 5-hour limits faster. A Sept 25 GitHub issue is titled: "Headless claude -p (entrypoint sdk-cli) uses ~1.8× more of the 5-hour window per token than interactive cli, same workload". Saturday at 9:25 PM ET, T3 Code's Theo posted: "All Agents SDK usage will drain your 5h limits faster than using Claude Code directly," measured 1.6x to 1.7x there, and wrote: "To be clear: this does not affect weekly limits." Anthropic hadn't replied publicly at our 5:52 AM ET check. (reported)
• Two overnight Claude error incidents, both resolved. Claude's status page logged two incidents early Sunday, each listing claude.ai, the Claude API, Claude Code and Claude Cowork. First: impact from 2:05 to 2:35 AM ET; it opened naming Mythos 5.1, Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 5.5, and the cause was identified on Sonnet 5.5 (status). Second: Fable 5.1, impact from 4:08 to 4:29 AM ET (status). Both are resolved, and the page read "All Systems Operational" at our 5:50 AM ET check. Anthropic hasn't posted a cause. (confirmed)
• Two Claude Code builders say they've left Anthropic. Thariq Shihipar posted at 12:04 AM ET Saturday: "I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined." Anthony Morris, whose bio says he created Claude Code on desktop, posted at 8:33 AM ET: "Yesterday was my last day at Anthropic." He added: "I'll share what's next soon!" Alex Albert, whose bio reads "Research @AnthropicAI," later listed staff to tag: "If you ever have feedback, tag them:" Neither departing post gives a reason. (confirmed)
• Unreleased: a "Harbor lantern" limit shows up in one user's usage tracker. Unreleased: developer Greg Schier posted a screenshot at 12:06 AM ET Sunday asking: "Claude new model leak in my usage tracker?" It shows a row named Harbor lantern at 0%, with its own reset timer, between Weekly and Weekly · Fable. Other posts we found repeat his screenshot. Anthropic hasn't confirmed what it is, and we found no second independent sighting. (rumor)

Hark#

• iOS ratings keep rising with no new version. Apple's lookup data for the Hark Pro listing showed 4.65 stars from 1,572 ratings at 6:06 AM ET Sunday, up from 4.63 from 1,446 at 6:00 AM ET Saturday. Our math: that is 126 new ratings in about a day, down from 386 the day before. The app is still version 2.4.0 from Oct 16, and the Google Play listing shows no new update. (store listings)
• Unverified: posts claim Hark Pro runs on Claude Opus 5.5. Several Chinese-language posts on X this weekend say Hark Pro is built on Anthropic's Opus 5.5, such as YB and Chsengni (user reports). Another user, 鲸鱼知识库, noted that Hark has never confirmed this and that it traces to a label a third party saw in a configuration, which may not be what actually runs. We found no Hark statement naming the model. Hark's March launch release says it is building its own foundation models. (rumor)