Jonathan Quinn
@jonoquinn.bsky.social
Reposted by Jonathan Quinn
It's a small contribution, but I finally bit the bullet and removed all references to Twitter/X from my website. Where previously I had embedded tweeprints on my website, I've now pasted all the content into something without links/mentions of twitter, e.g. here neural-reckoning.org/pub_multimod...
Nonlinear fusion is optimal for a wide class of multisensory tasks
Animals continuously detect information via multiple sensory channels, like vision and hearing, and integrate these signals to realise faster and more accurate ...
neural-reckoning.org
January 16, 2026 at 5:26 PM
It's a small contribution, but I finally bit the bullet and removed all references to Twitter/X from my website. Where previously I had embedded tweeprints on my website, I've now pasted all the content into something without links/mentions of twitter, e.g. here neural-reckoning.org/pub_multimod...
Reposted by Jonathan Quinn
Nobel Please Prize
January 15, 2026 at 10:56 PM
Nobel Please Prize
Reposted by Jonathan Quinn
Went back home & saw how my mom watched Chinese TV. A back to back YouTube AI slop of consciousness. I've never seen videos like this before. About teeth, Costco deals, healthy eating... no pauses & unhinged. "This is all just unchecked AI nonsense! Stop, mom!" I said as I snatched the remote away.
January 15, 2026 at 7:19 PM
Went back home & saw how my mom watched Chinese TV. A back to back YouTube AI slop of consciousness. I've never seen videos like this before. About teeth, Costco deals, healthy eating... no pauses & unhinged. "This is all just unchecked AI nonsense! Stop, mom!" I said as I snatched the remote away.
Reposted by Jonathan Quinn
"No one ever had or was a self. Buuuhhh, continuity of consciousness is just an illusion experienced by each and every one of a series of short-lived 'phenomenal selves.'"
Either erase my consciousness or give me my money back, you philosophy bastard.
Either erase my consciousness or give me my money back, you philosophy bastard.
January 15, 2026 at 3:19 AM
"No one ever had or was a self. Buuuhhh, continuity of consciousness is just an illusion experienced by each and every one of a series of short-lived 'phenomenal selves.'"
Either erase my consciousness or give me my money back, you philosophy bastard.
Either erase my consciousness or give me my money back, you philosophy bastard.
Reposted by Jonathan Quinn
#Measles cases continues to climb. The 2025 tally has now hit 2242 & there've already been more confirmed cases in 2026 — 171 — than were logged in most years from the late 1990s to 2023!
Interesting to see the age distribution of cases. Lots of susceptible adults, seems. www.cdc.gov/measles/data...
Interesting to see the age distribution of cases. Lots of susceptible adults, seems. www.cdc.gov/measles/data...
January 14, 2026 at 10:39 PM
#Measles cases continues to climb. The 2025 tally has now hit 2242 & there've already been more confirmed cases in 2026 — 171 — than were logged in most years from the late 1990s to 2023!
Interesting to see the age distribution of cases. Lots of susceptible adults, seems. www.cdc.gov/measles/data...
Interesting to see the age distribution of cases. Lots of susceptible adults, seems. www.cdc.gov/measles/data...
Reposted by Jonathan Quinn
ICYMI: My column for @chemistryworld.com on the challenges of definitions of life, and why I don't think replication is the big deal it's usually made out to be. Those are mules, by the way.
www.chemistryworld.com/opinion/arch...
www.chemistryworld.com/opinion/arch...
Archaeon’s lack of metabolism challenges definitions of life
A question that is not the same as asking whether something is alive
www.chemistryworld.com
January 14, 2026 at 9:01 AM
ICYMI: My column for @chemistryworld.com on the challenges of definitions of life, and why I don't think replication is the big deal it's usually made out to be. Those are mules, by the way.
www.chemistryworld.com/opinion/arch...
www.chemistryworld.com/opinion/arch...
Reposted by Jonathan Quinn
Odds ratio 2.06, and the methods look good to me. Not a huge deal but worth checking if you've downloaded your 23andme data or similar.
A specific genetic difference in a gene called RTP5—known to be involved in smell and taste perception—was linked to a higher risk of early death in patients hospitalized with severe COVID-19, according to a recent study led by researchers at Harvard Chan School. @harvardenvhealth.bsky.social
Genetic difference may help explain wide variation in COVID severity | Harvard T.H. Chan School of Public Health
A specific genetic difference in a gene called RTP5—known to be involved in smell and taste perception—was linked to a higher risk of early death in patients hospitalized with severe COVID-19, a study...
hsph.harvard.edu
January 13, 2026 at 4:58 AM
Odds ratio 2.06, and the methods look good to me. Not a huge deal but worth checking if you've downloaded your 23andme data or similar.
Reposted by Jonathan Quinn
AI writing detection is getting _really_ good these days. I hadn’t been paying attention to this part of the industry and I’m surprised (in a good way) to see how good it has gotten!
papers.ssrn.com/sol3/papers....
papers.ssrn.com/sol3/papers....
January 10, 2026 at 10:18 AM
AI writing detection is getting _really_ good these days. I hadn’t been paying attention to this part of the industry and I’m surprised (in a good way) to see how good it has gotten!
papers.ssrn.com/sol3/papers....
papers.ssrn.com/sol3/papers....
Reposted by Jonathan Quinn
The cost of sequencing a human genome has fallen over 100,000 fold in nominal terms since 2001.
In a new visualization, I've added some of the key advances in sequencing during that timeline:
In a new visualization, I've added some of the key advances in sequencing during that timeline:
January 8, 2026 at 8:16 AM
The cost of sequencing a human genome has fallen over 100,000 fold in nominal terms since 2001.
In a new visualization, I've added some of the key advances in sequencing during that timeline:
In a new visualization, I've added some of the key advances in sequencing during that timeline:
Reposted by Jonathan Quinn
Finally! Tens of thousands of asteroids are found each year. Yet 2026 had ZERO through Jan 6.
The drought is over.
Possibly as big as a jet liner. Possibly as large as a football field. Meet 2026 AA.
Blame the Full Moon for the drought.🌕 Expect many more soon.
The drought is over.
Possibly as big as a jet liner. Possibly as large as a football field. Meet 2026 AA.
Blame the Full Moon for the drought.🌕 Expect many more soon.
January 8, 2026 at 7:19 AM
Reposted by Jonathan Quinn
The must-read paper on LLMs, language, and thought that I reference here:
Dissociating language and thought in large language models
arxiv.org/abs/2301.06627
by @kmahowald.bsky.social @neuranna.bsky.social Idan Blank @nancykanwisher.bsky.social @joshtenenbaum.bsky.social @evfedorenko.bsky.social
Dissociating language and thought in large language models
arxiv.org/abs/2301.06627
by @kmahowald.bsky.social @neuranna.bsky.social Idan Blank @nancykanwisher.bsky.social @joshtenenbaum.bsky.social @evfedorenko.bsky.social
January 7, 2026 at 4:19 PM
The must-read paper on LLMs, language, and thought that I reference here:
Dissociating language and thought in large language models
arxiv.org/abs/2301.06627
by @kmahowald.bsky.social @neuranna.bsky.social Idan Blank @nancykanwisher.bsky.social @joshtenenbaum.bsky.social @evfedorenko.bsky.social
Dissociating language and thought in large language models
arxiv.org/abs/2301.06627
by @kmahowald.bsky.social @neuranna.bsky.social Idan Blank @nancykanwisher.bsky.social @joshtenenbaum.bsky.social @evfedorenko.bsky.social
Reposted by Jonathan Quinn
Anyone know any good papers on measuring performance of LLMs? Everything I've seen is pretty bad: full of confounds, biases, cherry picking, poor baselines, etc. I'm wondering if it's impossible for the same reason that measuring "IQ" is impossible, and is so if there is a reasonable thing to do.
January 7, 2026 at 9:38 AM
Anyone know any good papers on measuring performance of LLMs? Everything I've seen is pretty bad: full of confounds, biases, cherry picking, poor baselines, etc. I'm wondering if it's impossible for the same reason that measuring "IQ" is impossible, and is so if there is a reasonable thing to do.
Reposted by Jonathan Quinn
The resolution of microscopes has increased over 10,000 fold over the last 200 years.
It's allowed scientists to examine not only cells, but bacteria, then viruses, their protein structure, and, eventually, the individual atoms that comprise them.
It's allowed scientists to examine not only cells, but bacteria, then viruses, their protein structure, and, eventually, the individual atoms that comprise them.
January 7, 2026 at 4:42 PM
The resolution of microscopes has increased over 10,000 fold over the last 200 years.
It's allowed scientists to examine not only cells, but bacteria, then viruses, their protein structure, and, eventually, the individual atoms that comprise them.
It's allowed scientists to examine not only cells, but bacteria, then viruses, their protein structure, and, eventually, the individual atoms that comprise them.
Reposted by Jonathan Quinn
The proliferation of hummingbird feeders has become a “major evolutionary force” for the Anna’s hummingbird species in the western US. “Over just a few generations, their beaks have dramatically changed in size and shape.” [science.org]
Bird feeders have caused a dramatic evolution of California hummingbirds
Beaks have grown longer and larger, and ranges have expanded to follow the feeders
www.science.org
January 7, 2026 at 2:48 PM
The proliferation of hummingbird feeders has become a “major evolutionary force” for the Anna’s hummingbird species in the western US. “Over just a few generations, their beaks have dramatically changed in size and shape.” [science.org]
The central idea – that we can think of the genome as instantiating a generative model of the organism – is laid out in this recent paper with Nick Cheney: www.sciencedirect.com/science/arti...
January 6, 2026 at 9:43 AM
Reposted by Jonathan Quinn
Starting the day with a sunrise paddle
January 5, 2026 at 9:43 AM
Starting the day with a sunrise paddle
Reposted by Jonathan Quinn
For today’s Exhibit of the Day, let’s visit the Tarugo, a dazzling elbaite tourmaline. This gem is one of the finest large crystals ever found. It is notable for its exceptional quality and unique cranberry color.
January 5, 2026 at 2:28 AM
For today’s Exhibit of the Day, let’s visit the Tarugo, a dazzling elbaite tourmaline. This gem is one of the finest large crystals ever found. It is notable for its exceptional quality and unique cranberry color.
Reposted by Jonathan Quinn
Beautiful - recommended! Here, @sasolla.bsky.social recaps her decades-long journey from physics to neural networks (working with LeCun & Hopfield) to motor cortex, & and from industry (including Bell Labs) to academia, all driven by curiosity and awe (which flows from her voice). Inspiring!
Episode #36 in #TheoreticalNeurosciencePodcast: On low-dimensional manifolds in motor cortex – with Sara Solla @sasolla.bsky.social
theoreticalneuroscience.no/thn36
Manifold analysis has changed our thinking on how cortex works. One of the pioneers of this modelling approach explains.
theoreticalneuroscience.no/thn36
Manifold analysis has changed our thinking on how cortex works. One of the pioneers of this modelling approach explains.
January 4, 2026 at 11:03 AM
Beautiful - recommended! Here, @sasolla.bsky.social recaps her decades-long journey from physics to neural networks (working with LeCun & Hopfield) to motor cortex, & and from industry (including Bell Labs) to academia, all driven by curiosity and awe (which flows from her voice). Inspiring!
Reposted by Jonathan Quinn
Reposted by Jonathan Quinn
Fascinated by in-context learning (ICL) by LLMs recently.
In the Gemini 1.5 paper (old, I know) they tested ICL by asking the model to translate Kalamang (<200 speakers globally) to English.
With no context, it gives gibberish. With a Kalamang reference, it does really well.
In the Gemini 1.5 paper (old, I know) they tested ICL by asking the model to translate Kalamang (<200 speakers globally) to English.
With no context, it gives gibberish. With a Kalamang reference, it does really well.
January 4, 2026 at 7:25 AM
Fascinated by in-context learning (ICL) by LLMs recently.
In the Gemini 1.5 paper (old, I know) they tested ICL by asking the model to translate Kalamang (<200 speakers globally) to English.
With no context, it gives gibberish. With a Kalamang reference, it does really well.
In the Gemini 1.5 paper (old, I know) they tested ICL by asking the model to translate Kalamang (<200 speakers globally) to English.
With no context, it gives gibberish. With a Kalamang reference, it does really well.
Reposted by Jonathan Quinn
This is an artificial mouth created by Japanese researchers. It uses artificial vocal cords, a trachea system, and a flexible silicone mouth that moves and shapes sound in real time.
It sounds so human-like!
It sounds so human-like!
December 28, 2025 at 9:53 PM
This is an artificial mouth created by Japanese researchers. It uses artificial vocal cords, a trachea system, and a flexible silicone mouth that moves and shapes sound in real time.
It sounds so human-like!
It sounds so human-like!
Reposted by Jonathan Quinn
Autonomous RIVR delivery robots in Pittsburgh
December 24, 2025 at 6:56 PM
Autonomous RIVR delivery robots in Pittsburgh
Reposted by Jonathan Quinn
The Breakthrough Listen program conducted a technosignature search toward 3I/ATLAS using the 100 m Robert C. Byrd Green Bank Telescope at 1-12 GHz. We report a nondetection of candidate signals down to the 100 mW level. (Jacobson-Bell et al. submitted to RNAAS) arxiv.org/abs/2512.19763
Breakthrough Listen Observations of 3I/ATLAS with the Green Bank Telescope at 1-12 GHz
3I/ATLAS, an interstellar object, made its closest approach to Earth on 2025 December 19. On 2025 December 18, the Breakthrough Listen program conducted a technosignature search toward 3I/ATLAS using ...
arxiv.org
December 24, 2025 at 10:17 AM
The Breakthrough Listen program conducted a technosignature search toward 3I/ATLAS using the 100 m Robert C. Byrd Green Bank Telescope at 1-12 GHz. We report a nondetection of candidate signals down to the 100 mW level. (Jacobson-Bell et al. submitted to RNAAS) arxiv.org/abs/2512.19763
Reposted by Jonathan Quinn
The replies and QTs on this have the brainworm nature.
Scientists have developed a robotic hand exoskeleton that can passively train expert pianists to play faster keystrokes and overcome the ceiling effect in motor skills.
Read more in #ScienceRobotics: https://scim.ag/3N3bJil
Read more in #ScienceRobotics: https://scim.ag/3N3bJil
December 24, 2025 at 3:06 AM
The replies and QTs on this have the brainworm nature.
Reposted by Jonathan Quinn
Thanks @brianclegg.bsky.social for such a thoughtful review! Delighted that the book was of interest!
Review: The Giant Leap: @calebscharf.bsky.social **** - An invigorating mix of investigating the technology required to get humans into space and the idea that expanding into space is our evolutionary next step. popsciencebooks.blogspot.com/2025/12/the-... #bookreview #popularscience #spacetravel
The Giant Leap - Caleb Scharf ****
This is surely Caleb Scharf's most personal work - and certainly quite different from some of his earlier output, such as his excellent Grav...
popsciencebooks.blogspot.com
December 23, 2025 at 6:04 PM
Thanks @brianclegg.bsky.social for such a thoughtful review! Delighted that the book was of interest!