17 connected korrents · 14 moments from 18 Mar 2024 to 21 Oct 2026.
At the centre
Four things about a context window decide what an agent does next: its size, whether the information in it is correct, whether information is missing, and the trajectory it is already on.
Last stated 15 Jul 2026 · 3 months ago
Holds
Dex Horthy
Read this korrent →
-
Similar wording
A bigger context window does not give you a smarter model; the intelligence of the model is what decides how much of that window it can actually attend to.
Last stated 15 Jul 2026 · 3 months ago
Holds
Dex Horthy
-
Similar wording
Every announcement of a longer context window hides the cost, which is that the model gets worse at following instructions.
Last stated 19 Jun 2024 · 2 years ago
Holds
Aravind Srinivas
-
Similar wording
The point of context engineering is to do as much work as possible in the smart zone — roughly the first hundred thousand tokens of the context window.
Last stated 15 Jul 2026 · 3 months ago
Holds
Dex Horthy
-
Similar wording
If you will write a huge markdown file so an agent has context, you should have written the documentation that gives a human the same context.
Last stated 3 Jun 2026 · 5 months ago
Holds
Kelsey Hightower
-
Similar wording
Before an agent generates any UI for an exploration, have it describe the approaches in text, which is quicker to judge as directionally right or wrong.
Last stated 18 Aug 2026 · 2 months ago
Holds
Patrick Morgan
-
Similar wording
An agent repeats its own history: whatever it did on the last change — ran the tests or skipped them — is what it will do on the next one, because it is predicting the next message in the conversation.
Last stated 15 Jul 2026 · 3 months ago
Holds
Dex Horthy
-
Similar wording
How much someone needs to know depends on whether they want a fun demo or prototype or a product to sustain over time.
Last stated 25 Mar 2026 · 7 months ago
Holds
Flora Guo
-
Similar wording
Two agents are better than one: your agent can hand another agent far more context about you than you could ever type.
Last stated 24 May 2026 · 5 months ago
Holds
Dan Shipper
-
Same subject: Anthropic
A benchmark that ranks Claude Code last while it stays first in use is measuring the wrong thing, and has been for a year.
Last stated 3 Sept 2026 · 2 months ago
Holds
Dax Raad
-
Same subject: Anthropic
A skill is a set of repeatable, specific instructions that augments what Claude Code can do.
Last stated 19 Mar 2026 · 7 months ago
Holds
Kris Puckett
-
Same subject: Anthropic
A story's quality depends on whether it engages and satisfies the reader, not on the tool used to write it.
Last stated 30 Aug 2026 · 2 months ago
Holds
Eliot Peper
-
Same subject: OpenAI
A PHP vulnerability found with GPT-5.6 Sol will cause havoc this week.
Last stated 21 Oct 2026 · yesterday
Holds
Peter Steinberger
-
Same subject: OpenAI
A startup should not begin life as a nonprofit and bolt a for-profit arm on later, whatever OpenAI's own history suggests.
Last stated 18 Mar 2024 · 3 years ago
Holds
Sam Altman
-
Same subject: OpenAI
Advertising was a necessary phase for the internet but a momentary industry, and an AI people pay for is better because they know the answers are not influenced by advertisers.
Last stated 18 Mar 2024 · 3 years ago
Holds
Sam Altman
-
Same subject: developer tools
A compiler that also serves an IDE has to be lazy and deferred throughout — a different way of building compilers from the one the textbooks teach.
Last stated 13 May 2026 · 5 months ago
Holds
Anders Hejlsberg
-
Same subject: developer tools
A good tool raises the floor of what people can build without lowering the ceiling, and almost every tool quietly lowers the ceiling.
Last stated 18 Sept 2025 · a year ago
Holds
Tobias Lütke
-
Same subject: developer tools
A new API framework that ignores how existing apps were historically built undermines its own introduction
Last stated 31 Aug 2026 · 2 months ago
Holds
Craig Hockenberry