181.
182.
Even (very) noisy LLM evaluators are useful for improving AI agents
(news.ycombinator.com)
183.
Where does next-token prediction leave us?
(news.ycombinator.com)
184.
NotebookLM just made it easier to keep your sources up to date
(androidauthority.com)
185.
Barthelme, the Houstonian
(news.ycombinator.com)
186.
‘Obsession’ Just Did Something Most Movies Never Do
(gizmodo.com)
187.
189.
190.
A sleep-like consolidation mechanism for LLMs
(news.ycombinator.com)
191.
Eagle 3.1: Collaboration Between the EAGLE Team, vLLM Team, and TorchSpec Team
(news.ycombinator.com)
192.
12 pioneering phones that launched ahead of their time
(androidauthority.com)
193.
Investigating how prompt politeness affects LLM accuracy (2025)
(news.ycombinator.com)
194.
Prompt Politeness Affects LLM Accuracy
(news.ycombinator.com)
195.
A portentous reunion
(news.ycombinator.com)
196.
A Portentous Reunion
(news.ycombinator.com)
197.
‘James Bond’ Just Can’t Stop Learning to Be ‘James Bond’
(gizmodo.com)
199.
Reclaiming Social Engineering for Good
(spectrum.ieee.org)
200.
201.
Here’s How *That* ‘Obsession’ Scene Came Together
(gizmodo.com)
202.
Flick (YC F25) Is Hiring Front End Engineer to Build Figma for AI Filmmaking
(news.ycombinator.com)
203.
Constraint Decay: The Fragility of LLM Agents in Back End Code Generation
(news.ycombinator.com)
204.
205.
‘Mission: Impossible’ Has Always Been Important
(gizmodo.com)
206.
207.
The Amish Are Embracing ChatGPT
(futurism.com)
208.
--dangerously-skip-reading-code
(news.ycombinator.com)
209.
- -dangerously-skip-reading-code
(news.ycombinator.com)