OpenAI expands model lineup with GPT-5.6 variants and 1.5M token context window next week.
An assistant that reads your subscriptions so you don't have to. Point your newsletters at a dedicated alias; your assistant works in the background and forwards only what's worth your time. You'll hear from it when there's a reason to.
Filters route by sender. Summarizers condense everything. This reads each story and decides if it's for you.
Today's brief
OpenAI expands model lineup with GPT-5.6 variants and 1.5M token context window next week.
OpenAI's next model expands context window to 1.5M tokens with multiple tier variants.
OpenAI's next model iteration expands context window to 1.5M tokens with tiered variants.
Curated daily from 3 topics digestbox subscribes to. Pointer-with-attribution — every link goes to the publisher.
What are you actually trying to keep up with right now? Answer in a sentence or two. You can refine later.
Sign up to newsletters with [email protected] instead of your real inbox. Your real inbox stays untouched — only newsletters you point at the alias come through us.
Your assistant reads everything. You get one concise email per issue — or nothing at all.
how it reads
Your assistant reads every issue, so you don't.
It works in the background. No feed, no dashboard, no unread counter — just an inbox that stays quiet until something actually matters.
since launch
The funnel, end to end.
2,911 newsletters arrived. Inside: 5,074 stories. 784 cleared the bar — ≈ 35h 45m of skimming nobody had to do.
Newsletters → stories inside → stories that cleared the bar. Time saved estimated at 30 seconds per story headline + lede skimmed.