🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

We wrote up what we've learned about using Claude Code internally at Anthropic. Here are the most effective patterns we've found (many apply to coding with LLMs generally):

The US problems in this regard seem much greater than the British ones. I don't think this is a good time to mock Britain in this way. Poor judgement.

Someone just reminded me of this lecture I gave in 2009 that described the evolution of Google Search from 1999 to 2009. People who are interested in how our search systems work might find this interesting. It touches on disk-based serving systems, in-memory indices,

No we wont. Also, people seem to forget the 2021 expanded CTC was an increase of $23,600 over the childhood of a kid (or an increase of $57k if you were previously ineligible).

If you’re on that list of common defaults for dev solutions (React, Tailwind, Nextjs or Vite, Supabase, Playwright, Hasura, etc) then yeah, that’s going to be worth a lot

The "unilateral free trade" people are generally so oblivious to how it's so much more about addressing non tariff barriers vs getting tariff rates to zero in the 21st century

In case you haven’t heard, DeepL is now available as a default translator app on iPhones! Here’s how to set DeepL as your default, and make it even easier to access superior quality translations in the palm of your hand: ✅ Upgrade to iOS 18.4 ✅ Get the latest version of the

I didn't get to announce it since I had to give a major talk on this yesterday at Data Council, but Instant SQL is both out and it's on HN! Please clap (go vote for it and keep it up there) I'll do a proper post later today

in my new Tufts University Public Opinion Lab post, I confirm what dating app users already know - men (likely) lie. a lot. my data shows men overreport voting (+4 pts), blood donation (+3.5 pts), and even crime victimization (+1.4 pts). more here: tufts-pol.medium.com/no-you-are-not…

this is such a cool project…showing how you can make weighting adjustments to small subgroups in large sample surveys like the CES to get better estimates

New long post by me: The decline in cancer mortality is about much more than smoking. Many factors — including improvements in diagnosis, medical and surgical innovation, vaccination against cancers — have had a very real impact. scientificdiscovery.dev/p/the-decline-…

We've recently shipped a fresh batch of Claude Code features to improve your workflow! These updates make Claude Code faster, more visual, and easier to control. Here's what's new:

Value-added estimates for 14,000 nursing homes show substantial within-market heterogeneity, pointing to gains from policies that encourage patient reallocation. Liran Einav, Amy Finkelstein, Neale Mahoney econometricsociety.org/publications/e…

Announcing Datasette for Newsrooms - hosted Datasette specifically targeted at newsrooms and data journalists Think of it as a library for your data - load in CSVs and JSON, extract data with LLMs, collaborate on analysis with the rest of your team simonwillison.net/2025/Apr/24/in…

As someone who's spent a fair amount of time nerding out on embedded databases, this one by Semih Salihoğlu is one for the ages! "What every developer needs to know about in-processes DBMSs" Kùzu, LanceDB and DuckDB all feature! graphgeeks.org/blog/what-ever…

To piggyback off of this thread, Cremieux has no real opinions on psychometrics. Every take, and every (counter)argument he has about the Flynn effect, latent variables, group differences, and even heritability is taken almost verbatim from researchers he agrees with.

Today is the Last Day to get 35% off our evals course! Use this code: maven.com/parlance-labs/… We also have a very solid lineup of guest speakers, spanning topics in eval methodologies, observability tooling, pipeline architectures (e.g., RAG), and applications (e.g., voice).

My fav blog from Phil Eaton This kinda stuff motivates me to learn new things and try to see what I'm really interested in

Possibly, but do not reason on long-term tax elasticities based on short-term response that will not apply except for the period around the tax change itself.

Nice post. I also agree with the idea to always start with rigorous & systematic qualitative evals. There is a way to make vibe checks rigorous! And it’s necessary before scaling up!