K300 Ventures

Most governance tools live in their own consoles or dashboards. LLM Gateway lives in LangSmith. When requests are blocked or info is redacted, you get traceable events. See what your agents do, update system prompts or tool configs, re-evaluate against existing test sets, all in one place.

new team launching [REDACTED] on Push Chain today. any guesses?

Solana's favorite masochist doing legendary things. https://x.com/redacted_noah/status/2062342117293457776

Today we published a protocol update, providing the latest on where we are on relaunch and a path to user recovery. Excited to announce that @redacted_noah, one of Solana’s top smart contract engineers, will be joining Drift as Head of Protocol and previous members of the @gauntlet_xyz team contrib

迷因币四大天王：狗狗币、柴犬、佩佩、MemeCore，现在到底谁在C位领跑？🤔 我all in了狗子，评论区亮出你的持仓+战队，让我看看谁才是真·信仰玩家！ #迷因币 #加密货币

SAROS/USDT 180s Up 10.08% $0.000539 rose to $0.000593

$NVDA $INTC Reassuring to know on-net participants are still [redacted]. Not really sure how there is alpha in this as the amount of selling on these names into close was known several days ago. But clearly there was some interesting trades into the end of the day. Sheesh… https://x.com/markflow

Another reason why agentic search over transcripts is insufficient for recall: thinking blocks only contain summaries of thinking at best (they might be entirely redacted or encrypted). You actually need to *talk* to the transcript directly to access the precise thinking traces.

Before LangSmith LLM Gateway: An agent processes a request that includes a SSN. It now sits in LLM provider logs, in trace data, + possibly in downstream systems that consumed the response. With LangSmith LLM Gateway: Data is redacted from requests before it hits a model or gets logged to a trace

My brain before coffee: API Error: 400 `redacted_thinking` blocks in the latest assistant message cannot be modified