Jeffrey Barrett
@jeffbarrett.eu
USA->UK->Finland. Posting about 'omics technologies and the great outdoors (among other things).
I am presenting updates on multiomic research and clinical risk prediction on behalf of @nightingalehealth.com at this year's @ukbiobank.bsky.social conference. Looking forward to joining the event, and catching up with other attendees!
#UKBiobank #Omics #HealthData
#UKBiobank #Omics #HealthData
October 31, 2025 at 7:37 AM
I am presenting updates on multiomic research and clinical risk prediction on behalf of @nightingalehealth.com at this year's @ukbiobank.bsky.social conference. Looking forward to joining the event, and catching up with other attendees!
#UKBiobank #Omics #HealthData
#UKBiobank #Omics #HealthData
This part of @billkristolbulwark.bsky.social ‘s piece is spot on. Another reason “No Kings” is better is because it’s a good Saxon word, whereas the others are effete Latin words.
www.orwellfoundation.com/the-orwell-f...
www.orwellfoundation.com/the-orwell-f...
October 22, 2025 at 4:55 PM
This part of @billkristolbulwark.bsky.social ‘s piece is spot on. Another reason “No Kings” is better is because it’s a good Saxon word, whereas the others are effete Latin words.
www.orwellfoundation.com/the-orwell-f...
www.orwellfoundation.com/the-orwell-f...
Saturday at ASHG is always bittersweet. Still listening to cool science, but many people have gone home, and the “see you next year” signs are up. It goes by fast! #ashg25
October 18, 2025 at 2:07 PM
Saturday at ASHG is always bittersweet. Still listening to cool science, but many people have gone home, and the “see you next year” signs are up. It goes by fast! #ashg25
“Hey Frank, we need some art for the lunar eclipse story, stat!”
“I’ve got just the thing, boss.”
“I’ve got just the thing, boss.”
September 7, 2025 at 3:21 PM
“Hey Frank, we need some art for the lunar eclipse story, stat!”
“I’ve got just the thing, boss.”
“I’ve got just the thing, boss.”
The bluesky ecosystem: the menswear guy being quote posted by a prof. of political science, being reposted to my timeline by an astronomer.
August 26, 2025 at 11:49 AM
The bluesky ecosystem: the menswear guy being quote posted by a prof. of political science, being reposted to my timeline by an astronomer.
GPT5 fast mode sucks at things I expect it to be good at. Here's an example where it seemingly realizes it's full of shit halfway through its answer, and tries to recover:
August 11, 2025 at 3:26 PM
GPT5 fast mode sucks at things I expect it to be good at. Here's an example where it seemingly realizes it's full of shit halfway through its answer, and tries to recover:
This is so dumb, but I'm kind of addicted to this game.
August 8, 2025 at 12:06 PM
This is so dumb, but I'm kind of addicted to this game.
Do you think this is really still a tokenization limitation, or an OpenAI easter egg?
August 8, 2025 at 5:48 AM
Do you think this is really still a tokenization limitation, or an OpenAI easter egg?
Don't mess with the OG.
August 7, 2025 at 1:38 PM
Don't mess with the OG.
Here's a fun bit of math in R:
August 1, 2025 at 9:17 AM
Here's a fun bit of math in R:
As you both know, I bring a propane stove to all international meetings.
May 24, 2025 at 3:50 PM
As you both know, I bring a propane stove to all international meetings.
May 23, 2025 at 12:50 PM
20 years ago today, at 5:50 AM, a little more than a week after I had arrived in Oxford, I was headed to see the Magdalen choir. Pretty magical introduction to my new home!
May 1, 2025 at 9:17 AM
20 years ago today, at 5:50 AM, a little more than a week after I had arrived in Oxford, I was headed to see the Magdalen choir. Pretty magical introduction to my new home!
The very good @338canada.bsky.social model prediction of the upcoming election over time is informative about how to tune Bayesian models with limited history. It was absolutely certain of a conservative victory for >1 year. But strange things do happen.
April 2, 2025 at 4:32 PM
The very good @338canada.bsky.social model prediction of the upcoming election over time is informative about how to tune Bayesian models with limited history. It was absolutely certain of a conservative victory for >1 year. But strange things do happen.
I opened a signal account in case any governments wanted to add me to group chats. I seem to be on the right track?
April 2, 2025 at 10:44 AM
I opened a signal account in case any governments wanted to add me to group chats. I seem to be on the right track?
I think it come in second behind “put only one figure on each slide”
March 12, 2025 at 4:40 PM
I think it come in second behind “put only one figure on each slide”
The point in the day when big winter boots feel like the comfiest slippers.
March 2, 2025 at 11:36 AM
The point in the day when big winter boots feel like the comfiest slippers.
There’s nothing more relaxing than reading the Finnish national broadcaster’s English news. They assume you can get international news in English somewhere else, so just cover the most important happenings in Finland.
February 23, 2025 at 6:19 PM
There’s nothing more relaxing than reading the Finnish national broadcaster’s English news. They assume you can get international news in English somewhere else, so just cover the most important happenings in Finland.
President Stubb doesn’t mess around: his visit today to one of the brigades on Finland’s border with Russia clearly contrasts with some other Presidents I could mention.
February 19, 2025 at 8:31 PM
President Stubb doesn’t mess around: his visit today to one of the brigades on Finland’s border with Russia clearly contrasts with some other Presidents I could mention.
This will be my only comment on the events of the day.
January 20, 2025 at 5:30 PM
This will be my only comment on the events of the day.
OMG look at the time series results from the VRTX pain trial. Hard to get such a close match if you gave everyone the same drug! Do they really believe the effect of the drug was perfectly cancelled out by site heterogeneity?
December 20, 2024 at 12:48 PM
OMG look at the time series results from the VRTX pain trial. Hard to get such a close match if you gave everyone the same drug! Do they really believe the effect of the drug was perfectly cancelled out by site heterogeneity?