All existing proposals are wrong because they are all right
Tynan Is Typing | Tynan Purdy | Offprint
Take a peek inside my brain. Internet thoughts. Design takes. Pontificating on life.
Tynan Is Typing | Tynan Purdy | Offprint
Take a peek inside my brain. Internet thoughts. Design takes. Pontificating on life.
Tynan Is Typing | Tynan Purdy | Offprint
Take a peek inside my brain. Internet thoughts. Design takes. Pontificating on life.
Issue 015, The June Edition. | Offprint
Get some accountability friends for your side projects, manage your typefaces, and listen to some non-AI generated jazz.
+ the art of footnotes | Are.na
expanded to include margin annotations and other multi-channel approaches towards text, information and credit.
FOSDEM 2026 - FOSDEM 2026 Call for Participation
Proposals for developer rooms and main track talks for FOSDEM 2026 can now be submitted!
It's working - Tynan's Leaflets
5 pieces of context on a single post, proving what the social web offers today
How Basecamp takes a feature from idea to completion
This is a behind the scenes look at how we work at Basecamp. It is Part II of my article, Getting from ⭐️⭐️ to ⭐️⭐️⭐️⭐️⭐️ In Part 1, I wrote about how we were unhappy with the App Store ratings of …
SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge
We introduce SimpleQA Verified, a 1,000-prompt benchmark for evaluating Large Language Model (LLM) short-form factuality based on OpenAI's SimpleQA. It addresses critical limitations in OpenAI's benchmark, including noisy and incorrect labels, topical biases, and question redundancy. SimpleQA Verified was created through a rigorous multi-stage filtering process involving de-duplication, topic balancing, and source reconciliation to produce a more reliable and challenging evaluation set, alongside improvements in the autorater prompt. On this new benchmark, Gemini 2.5 Pro achieves a state-of-the-art F1-score of 55.6, outperforming other frontier models, including GPT-5. This work provides the research community with a higher-fidelity tool to track genuine progress in parametric model factuality and to mitigate hallucinations. The benchmark dataset, evaluation code, and leaderboard are available at: https://www.kaggle.com/benchmarks/deepmind/simpleqa-verified.
SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge
We introduce SimpleQA Verified, a 1,000-prompt benchmark for evaluating Large Language Model (LLM) short-form factuality based on OpenAI's SimpleQA. It addresses critical limitations in OpenAI's benchmark, including noisy and incorrect labels, topical biases, and question redundancy. SimpleQA Verified was created through a rigorous multi-stage filtering process involving de-duplication, topic balancing, and source reconciliation to produce a more reliable and challenging evaluation set, alongside improvements in the autorater prompt. On this new benchmark, Gemini 2.5 Pro achieves a state-of-the-art F1-score of 55.6, outperforming other frontier models, including GPT-5. This work provides the research community with a higher-fidelity tool to track genuine progress in parametric model factuality and to mitigate hallucinations. The benchmark dataset, evaluation code, and leaderboard are available at: https://www.kaggle.com/benchmarks/deepmind/simpleqa-verified.
A/B/U Review System - ORI Onboarding
This is the rating system we use at ORI. Given a piece of information, rate it according to the following rubric:
Tynan is an incredible communicator and I couldn't recommend him more
Alana Purdy’s Husband ⁂
For every Atmosphere project or initiative, I offer copywriting services. Comms are hard, especially for our movement which is quite technical in nature sometimes. As a designer, my trade is communication. The right words inspire people and share the hope we all have for the internet.
I’m deeply grateful for what Tynan does. You should hit him up!
Alana Purdy’s Husband ⁂
For every Atmosphere project or initiative, I offer copywriting services. Comms are hard, especially for our movement which is quite technical in nature sometimes. As a designer, my trade is communication. The right words inspire people and share the hope we all have for the internet.
Don't worry guys, Part 2 is here: Houston, we've got a book & review lexicon! I don't think it is in it's final shape just yet, but I believe it to be a strong proposal, and well justified if I may say so, myself 😌
Road to a book lexicon (Part 2)
nick-the-sick.pckt.blog
Nick The Sick
I've been thinking a lot about how to define a shared book lexicon for the Lexicon Community Working Group: discourse.atmosphere.community/t/in-which-a-book-lexicon-is-… But, instead wrote a treatise against ISBNs. I hope I can link to it whenever someone asks why their import didn't work on @bookhive.buzz 😝
#atproto is there anything we can do to solve the issues plaguing the lexicon specification right now I don't think it's good that the reference TS and Go SDK do not agree with each other on how lexicons should be interpreted
Ricardo J. Méndez
Hmm. That's generated with prototypey and goat lex lint didn't complain, but I'll look into it, thanks.
