WSL3 Performance is about 5-60% faster than WSL2 depending on the workload
Commenters discuss switching from WSL2 to Linux and Microsoft's lack of care for its OS.
30 stories
Get a notification every evening
You hear about the digest when it is ready. Turn it off any time.
Notifications are on
You will hear about each evening’s digest when it is ready.
To get notifications on iPhone or iPad, add this site to your Home Screen first: tap Share, then Add to Home Screen. Open it from there to turn notifications on.
Notifications are blocked in this browser. Allow them in site settings and reload.
New notification sign-ups are closed for now. You can follow along with RSS.
Notifications could not be turned on. Check your connection and try again.
Notifications could not be turned off and are still on. Check your connection and try again.
The day’s top discussions
Commenters discuss switching from WSL2 to Linux and Microsoft's lack of care for its OS.
Commenters discuss losing Knuth's reward check and an error in his book.
Commenters discuss the interactive nature of Triple-A Minesweeper and the changes to the original Minesweeper.
One commenter compares the high Claude token cost to US hospital bills, while another praises the system's use of metaphor for information encoding.
Commenters note the model is based on Qwen and that Microsoft is moving towards local inference.
Commenters discuss that Valen's borrow checking could be useful in code-driven scenarios, but that Rust's "shared-xor-mutable" restriction is actually a key feature.
Commenters highlight the difficulty of overcoming one's own biases when writing READMEs and the importance of explaining what the software does.
Commenters attribute Taiwan's low obesity rate to meals prepared with fresh ingredients and recommend a book on ultraprocessed foods.
Commenters note that Anthropic's AI model submitted a false tip, but the impact was limited thanks to the police spam filter.
Commenters point out issues with using tilde in scripts and the unnecessary quoting of strings in Bash.
Commenters discuss token-free LLM research and that subword tokenizers peeked at future bytes.
Commenters suggest a C tutorial should emphasize compiler warnings and building abstractions with types.
Commenters discuss developing a similar tool and changes in their code review processes.
A commenter questions why archive.is is the most reliable and what it does differently.
A commenter states they prefer the 90s ICQ interface over modern chat apps.
A commenter confirms the 1TB of data usage was from local network scans and that Keurig sells the data to advertisers.
A commenter states that unikernels lack observability tools and that the article undersells the benefits of a full OS.
A commenter notes that DuckDB 2.0 improves recursive CTE performance and that there is an extension for graph data.
A commenter appreciates the article's exposition of applying diffusion models to a time series that is neither continuous nor discrete.
A commenter claims that Google does not screen ads and profits from fraud by allowing misleading ads.
A commenter calls this type of aggressive audio processing 'Bose barf' and finds it grating.
A commenter notes the company lacks a strong product moat but stands out due to good engineering and marketing teams.
A commenter reports the research claims OpenAI's Lean formal proof does not correspond to the natural language proof of Navier-Stokes equations.
A commenter questions Nvidia's strategy in acquiring Reflection AI, which it recently invested in, while having its own models.
One commenter argues all problems are solved with methods like sub-agents, while another says agents still fail on complex tasks.
Commenters note the drop in Yandex bot traffic is specific to Yandex and barely noticeable globally.
A smaller, more technical community
A commenter explains the "shredding" method of copying JSON fields into separate columns and Logfire's dynamic shredding approach.
Commenters discuss the definition of 'vectorized' in the Python ecosystem.