carbocation

I'm very confused about why they are showing an AUC for the relationship of a continuous score with a continuous phenotype.

It originated in a comment on a blog: https://crookedtimber.org/2018/03/21/liberals-against-progre...

The public overwhelmingly supports NIH-funded research. https://www.aacr.org/about-the-aacr/newsroom/news-releases/n...

That is essentially how it works in the US as well thanks to public service loan forgiveness for physicians.

This scans very much as AI-written.

Was the choice of such a small model driven by a desire for high tok/sec? I ask because an m4 pro 48gb machine can run larger models (if model intelligence is the thing that would make it more useful).

Last month, Steve Yegge suggested that they are not: https://xcancel.com/Steve_Yegge/status/2043747998740689171

Neat, thank you for explaining!

Is there a page where I could read more? What's unintuitive at a glance is that Opus 4.7 has a lower success rate than Sonnet 4.6 (90% vs 100%) while having a higher Avg Percentile (87.2% vs 70.9%).

The $20 Plus plan still exists, and does not give access to the pro model.

The $200 Pro plan still exists, and does give access to the pro model.

What is new is a $100 Pro plan that does give access to the pro model, with lower usage limits than the $200 Pro plan.

It would be very helpful if the chat logs could (optionally) be retained.

The article kind of downplays the most interesting elements. Not an expert, but to my limited understanding:

* I think this is the longest-range use of a ballistic missile in anger, possibly ever?

* This seems to reveal previously-unknown range of Iranian ballistic missiles and, if true, could touch basically all of Europe?

I think the average is ballpark 1 per 100 million miles, so my guess is that someday they might show this, but they probably want to wait til they have more miles driven.

Haber-Bosch process, for the curious.

Because many of us think simonw has discerning taste on this topic and like to read what he has to say about it, so we upvote his comments.

I agree with this; it feels like the most likely tool to drop its high-level comments in code comments.

This model was able to transcribe Bad Bunny lyrics over the sound of the background music, played casually from my speakers. Impressive, to me.

300 grams of oatmeal a day, basically nothing else, and your LDL only goes down 10%.

Em-dashes are easy to type on a macos laptop for what it's worth: option-shift-minus.

The erdosproblems thread itself contains comments from Terence Tao: https://www.erdosproblems.com/forum/thread/281

That repo is throwing up a 404 for me.

Question - did you consider tradeoffs between duckdb (or other columnar stores) and SQLite?

The article explicitly says that the aircraft is safe. I don't think this is particularly clickbait-y.

> The public database of reported post-vaccination health issues is often misused to sow misinformation.

https://publichealth.jhu.edu/2022/what-vaers-is-and-isnt

VAERS cannot be used to establish causality; it cannot correctly be used in the way in which they are purporting to use it[1].

1 = https://www.kff.org/quick-take/fda-memo-linking-covid-vaccin...

If the URL is public, it would be neat to see if you want to share it.

It would be great to have access to this model via the chat interface, even if it was gated behind the "other models" dropdown or something.

To add more information, the intervention was guidance about caffeine intake. From the Methods:

> If allocated to caffeinated coffee consumption, patients were encouraged to drink at least 1 cup of caffeinated coffee (or at least 1 espresso shot) and other caffeine-containing products every day as per their usual lifestyle. It was recommended that patients in the coffee consumption group not intentionally increase or decrease consumption of coffee or other caffeine-containing products.

> If allocated to the abstinence group, patients were encouraged to completely abstain from coffee, including decaffeinated coffee, and other caffeine containing products.

Waymo was on a roll in San Francisco. It still is, but it used to be, too. (With apologies to Mitch.) This is utter sensationalism. Fortunately, the state's regulations have liberated the good people of SF from being able to shoot ourselves in our own foot on this particular issue.