Dominik Steiger
@dominiksteiger.bsky.social
Interested in politics and policy, science and health, the analogue going digital, and endemic delusion. Views my own.
Interesting preprint points to the problem of AI-caused slopification of the training sets for generative AI in medicine, absent human oversight.
www.medrxiv.org/content/10.6...
www.medrxiv.org/content/10.6...
February 6, 2026 at 7:51 AM
Interesting preprint points to the problem of AI-caused slopification of the training sets for generative AI in medicine, absent human oversight.
www.medrxiv.org/content/10.6...
www.medrxiv.org/content/10.6...
Reposted by Dominik Steiger
Nice to see that German specimen sequencing shows a similar dynamic, with BA.3.2* peaking in calendar week 51.
bsky.app/profile/rv-e...
bsky.app/profile/rv-e...
February 5, 2026 at 5:48 PM
Nice to see that German specimen sequencing shows a similar dynamic, with BA.3.2* peaking in calendar week 51.
bsky.app/profile/rv-e...
bsky.app/profile/rv-e...
#Covid
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 28. Januar 2026 (nur ein neuer Datenpunkt für Zürich):
Inzidenz ~250.
Ca. jede 400. Person infiziert sich pro Woche.
bsky.app/profile/domi...
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 28. Januar 2026 (nur ein neuer Datenpunkt für Zürich):
Inzidenz ~250.
Ca. jede 400. Person infiziert sich pro Woche.
bsky.app/profile/domi...
February 5, 2026 at 10:06 AM
#Covid
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 28. Januar 2026 (nur ein neuer Datenpunkt für Zürich):
Inzidenz ~250.
Ca. jede 400. Person infiziert sich pro Woche.
bsky.app/profile/domi...
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 28. Januar 2026 (nur ein neuer Datenpunkt für Zürich):
Inzidenz ~250.
Ca. jede 400. Person infiziert sich pro Woche.
bsky.app/profile/domi...
#Individualbesteuerung ist auch eine Umverteilungsvorlage. Umverteilung von arm zu reich.
Quelle: www.tagesanzeiger.ch/individualbe...
bsky.app/profile/domi...
Quelle: www.tagesanzeiger.ch/individualbe...
bsky.app/profile/domi...
February 5, 2026 at 9:16 AM
#Individualbesteuerung ist auch eine Umverteilungsvorlage. Umverteilung von arm zu reich.
Quelle: www.tagesanzeiger.ch/individualbe...
bsky.app/profile/domi...
Quelle: www.tagesanzeiger.ch/individualbe...
bsky.app/profile/domi...
Manche in der SP merken, dass die vorgeschlagene Individualbesteuerung sozial ungerecht ist.
www.blick.ch/politik/ring...
www.blick.ch/politik/ring...
Individualsteuer: Prominente SP-Köpfe gegen eigene Partei
In der SP regt sich Widerstand gegen die Individualbesteuerung. Prominente Köpfe prangern Ungerechtigkeiten an und zweifeln daran, dass damit mehr Gleichberechtigung zwischen Mann und Frau erreicht wi...
www.blick.ch
February 5, 2026 at 7:00 AM
Manche in der SP merken, dass die vorgeschlagene Individualbesteuerung sozial ungerecht ist.
www.blick.ch/politik/ring...
www.blick.ch/politik/ring...
Die "saisonale" Covid-Impfung bleibt effektiv - und das ist Effektivität über bestehende Immunität (durch Infektionen oder vorbestehende Impfungen) hinaus.
Es ist nicht zu rechtfertigen, dass sie anders als Grippeimpfungen behandelt wird.
www.cidrap.umn.edu/covid-19/202...
Es ist nicht zu rechtfertigen, dass sie anders als Grippeimpfungen behandelt wird.
www.cidrap.umn.edu/covid-19/202...
2024-25 COVID vaccine 80% effective against death, CDC estimates
www.cidrap.umn.edu
February 5, 2026 at 6:51 AM
Die "saisonale" Covid-Impfung bleibt effektiv - und das ist Effektivität über bestehende Immunität (durch Infektionen oder vorbestehende Impfungen) hinaus.
Es ist nicht zu rechtfertigen, dass sie anders als Grippeimpfungen behandelt wird.
www.cidrap.umn.edu/covid-19/202...
Es ist nicht zu rechtfertigen, dass sie anders als Grippeimpfungen behandelt wird.
www.cidrap.umn.edu/covid-19/202...
Deconvolution for BA.3.2* is finally publicly available for Swiss wastewater (cyan curve). Seq data until Jan 11, viral load data until Jan 25.
It started growing beginning of December and run into trouble.
Also noticeable: NB.1.8.1* subvariant fraction seems to rise.
bsky.app/profile/domi...
It started growing beginning of December and run into trouble.
Also noticeable: NB.1.8.1* subvariant fraction seems to rise.
bsky.app/profile/domi...
February 4, 2026 at 8:06 AM
Deconvolution for BA.3.2* is finally publicly available for Swiss wastewater (cyan curve). Seq data until Jan 11, viral load data until Jan 25.
It started growing beginning of December and run into trouble.
Also noticeable: NB.1.8.1* subvariant fraction seems to rise.
bsky.app/profile/domi...
It started growing beginning of December and run into trouble.
Also noticeable: NB.1.8.1* subvariant fraction seems to rise.
bsky.app/profile/domi...
Brilliant
Daddy Trump has needs (new short) youtube.com/shorts/hCw8C...
Daddy Trump Has Needs
YouTube video by Aamon Animations
youtube.com
February 3, 2026 at 5:49 PM
Brilliant
Reposted by Dominik Steiger
Singapore, data until week 4 2026
Very high ARI numbers, now even more detached from Covid and influenza. What is it?
Covid stays very low, despite a new variant (RF.5) becoming dominant.
Very high ARI numbers, now even more detached from Covid and influenza. What is it?
Covid stays very low, despite a new variant (RF.5) becoming dominant.
February 3, 2026 at 4:49 PM
Singapore, data until week 4 2026
Very high ARI numbers, now even more detached from Covid and influenza. What is it?
Covid stays very low, despite a new variant (RF.5) becoming dominant.
Very high ARI numbers, now even more detached from Covid and influenza. What is it?
Covid stays very low, despite a new variant (RF.5) becoming dominant.
Excess mortality (65+) in Switzerland.
Very high excess mortality continues into week 4. Eastern Switzerland reports numbers higher than ever since the 2nd Covid wave.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
Very high excess mortality continues into week 4. Eastern Switzerland reports numbers higher than ever since the 2nd Covid wave.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
February 3, 2026 at 3:46 PM
Excess mortality (65+) in Switzerland.
Very high excess mortality continues into week 4. Eastern Switzerland reports numbers higher than ever since the 2nd Covid wave.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
Very high excess mortality continues into week 4. Eastern Switzerland reports numbers higher than ever since the 2nd Covid wave.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
The Netherlands, calendar week 4, ending Jan 26th:
Both sentinel and wastewater show a local minimum.
BA.3.2* was at ca. 50% according to sparse specimen sequencing up to week 2. Despite fast fraction growth, it could not counteract overall epidemiological drivers, so far.
Both sentinel and wastewater show a local minimum.
BA.3.2* was at ca. 50% according to sparse specimen sequencing up to week 2. Despite fast fraction growth, it could not counteract overall epidemiological drivers, so far.
February 3, 2026 at 12:33 PM
The Netherlands, calendar week 4, ending Jan 26th:
Both sentinel and wastewater show a local minimum.
BA.3.2* was at ca. 50% according to sparse specimen sequencing up to week 2. Despite fast fraction growth, it could not counteract overall epidemiological drivers, so far.
Both sentinel and wastewater show a local minimum.
BA.3.2* was at ca. 50% according to sparse specimen sequencing up to week 2. Despite fast fraction growth, it could not counteract overall epidemiological drivers, so far.
Nimmt man eine Dunkelziffer von 80 an, laufen die deutschen Inzidenzen gemäss Meldungen nach IfSG ähnlich wie die RKI-Grippeweb-Inzidenzen und die Schweizer Sentinel-Detektionen.
Die Viruslast gemäss AMELAG läuft seit Frühling 2024 zu tief (kumulative Inzidenz ca. Faktor 2 zu tief).
Die Viruslast gemäss AMELAG läuft seit Frühling 2024 zu tief (kumulative Inzidenz ca. Faktor 2 zu tief).
February 3, 2026 at 9:51 AM
Nimmt man eine Dunkelziffer von 80 an, laufen die deutschen Inzidenzen gemäss Meldungen nach IfSG ähnlich wie die RKI-Grippeweb-Inzidenzen und die Schweizer Sentinel-Detektionen.
Die Viruslast gemäss AMELAG läuft seit Frühling 2024 zu tief (kumulative Inzidenz ca. Faktor 2 zu tief).
Die Viruslast gemäss AMELAG läuft seit Frühling 2024 zu tief (kumulative Inzidenz ca. Faktor 2 zu tief).
Who would have thought the platform and American hegemony going to shit so completely within fifteen years.
#MyXAnniversary
#MyXAnniversary
February 2, 2026 at 9:28 AM
Who would have thought the platform and American hegemony going to shit so completely within fifteen years.
#MyXAnniversary
#MyXAnniversary
Individualbesteuerung: Ich gehe davon aus, dass es dann individuelle Prämienvergünstigung und individuelle Sozialhilfe gibt?
February 2, 2026 at 7:07 AM
Individualbesteuerung: Ich gehe davon aus, dass es dann individuelle Prämienvergünstigung und individuelle Sozialhilfe gibt?
Here is an obituary for the mortality displacement argument that is sometimes brought forward. It's just not there, notwithstanding wishful thinking.
jamanetwork.com/journals/jam...
jamanetwork.com/journals/jam...
February 1, 2026 at 2:58 PM
Here is an obituary for the mortality displacement argument that is sometimes brought forward. It's just not there, notwithstanding wishful thinking.
jamanetwork.com/journals/jam...
jamanetwork.com/journals/jam...
Reposted by Dominik Steiger
January 30, 2026 at 8:52 AM
I posted funny failures of LLMs, but I want to mention a few cases that really impressed me:
Example 1:
It translates Russian lyrics in a way that is impressive to me. I can discuss with it some interpretation or grammar questions. It generally translates well when prompted carefully.
Example 1:
It translates Russian lyrics in a way that is impressive to me. I can discuss with it some interpretation or grammar questions. It generally translates well when prompted carefully.
January 29, 2026 at 4:41 PM
I posted funny failures of LLMs, but I want to mention a few cases that really impressed me:
Example 1:
It translates Russian lyrics in a way that is impressive to me. I can discuss with it some interpretation or grammar questions. It generally translates well when prompted carefully.
Example 1:
It translates Russian lyrics in a way that is impressive to me. I can discuss with it some interpretation or grammar questions. It generally translates well when prompted carefully.
#Covid
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 25. Januar 2026:
Inzidenz ~300.
Ca. jede 330. Person infiziert sich pro Woche.
bsky.app/profile/domi...
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 25. Januar 2026:
Inzidenz ~300.
Ca. jede 330. Person infiziert sich pro Woche.
bsky.app/profile/domi...
January 28, 2026 at 8:20 PM
#Covid
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 25. Januar 2026:
Inzidenz ~300.
Ca. jede 330. Person infiziert sich pro Woche.
bsky.app/profile/domi...
Schätzung Infektionsinzidenzen für Zürich aufgrund Kalibrierung der Abwasserprävalenzen Werdhölzli, Daten bis 25. Januar 2026:
Inzidenz ~300.
Ca. jede 330. Person infiziert sich pro Woche.
bsky.app/profile/domi...
Nate Silver made a nice graph showing the self- respectively Bezos-inflicted downfall of the Washington Post.
Just a quick newsletter today on the sad state of the Washington Post. This chart tells the story.
January 28, 2026 at 5:22 PM
Nate Silver made a nice graph showing the self- respectively Bezos-inflicted downfall of the Washington Post.
"Wohin steuert Amerika? Aussenpolitisch ist Kamala Harris das grössere Risiko als Trump"
January 28, 2026 at 11:18 AM
"Wohin steuert Amerika? Aussenpolitisch ist Kamala Harris das grössere Risiko als Trump"
Excess mortality (65+) in Switzerland.
Very high excess mortality in calendar week 3, ending Jan 18th. Excess mortality peak is close to the peak of influenza A H3N2 viral load in the wastewater.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
Very high excess mortality in calendar week 3, ending Jan 18th. Excess mortality peak is close to the peak of influenza A H3N2 viral load in the wastewater.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
January 27, 2026 at 1:22 PM
Excess mortality (65+) in Switzerland.
Very high excess mortality in calendar week 3, ending Jan 18th. Excess mortality peak is close to the peak of influenza A H3N2 viral load in the wastewater.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
Very high excess mortality in calendar week 3, ending Jan 18th. Excess mortality peak is close to the peak of influenza A H3N2 viral load in the wastewater.
Newest weeks are extrapolated, so these values can still change.
bsky.app/profile/domi...
Reposted by Dominik Steiger
Singapore, data until week 3 2026
High ARI numbers, despite flu coming down and Covid very low, a pattern never seen since I follow. Which viruses cause this?
High ARI numbers, despite flu coming down and Covid very low, a pattern never seen since I follow. Which viruses cause this?
January 25, 2026 at 3:52 PM
Singapore, data until week 3 2026
High ARI numbers, despite flu coming down and Covid very low, a pattern never seen since I follow. Which viruses cause this?
High ARI numbers, despite flu coming down and Covid very low, a pattern never seen since I follow. Which viruses cause this?
Absolutely no signs of a BA.3.2* wave, neither in the Netherlands, nor in Germany (where it was erroneously invoked to explain growth in December), nor in Switzerland. Maybe some time in the future, but not now.
January 23, 2026 at 11:08 AM
Absolutely no signs of a BA.3.2* wave, neither in the Netherlands, nor in Germany (where it was erroneously invoked to explain growth in December), nor in Switzerland. Maybe some time in the future, but not now.
"And the fact that Trump may have pulled back for now doesn’t change the fact there’s been a fundamental rupture in the post–World War II global order. The spectacle of the United States behaving like an out-and-out mob boss can’t be unseen."
www.thebulwark.com/p/old-europe...
www.thebulwark.com/p/old-europe...
Old Europe. New Backbone.
A lesson in standing up to bullies from across the pond.
www.thebulwark.com
January 22, 2026 at 2:38 PM
"And the fact that Trump may have pulled back for now doesn’t change the fact there’s been a fundamental rupture in the post–World War II global order. The spectacle of the United States behaving like an out-and-out mob boss can’t be unseen."
www.thebulwark.com/p/old-europe...
www.thebulwark.com/p/old-europe...
Reposted by Dominik Steiger
January 22, 2026 at 7:46 AM