Matt Mack
@matthras.bsky.social
Hard-of-hearing + mild AuDHD mathematician, cellist, anime+manga+VTuber fan with a giant backlog of games and books to get through. Chronically online: [email protected], [email protected], matthras.com
Pinned
Matt Mack
@matthras.bsky.social
· Oct 23
Howdy! I'm currently a PhD student #mathbio but have an eclectic mix of interests/specialties including:
🧶knitting/crochet/cross-stitch
🦻accessibility esp. in maths
👨🏫maths education
🧮operations research
🎵classical music
And am always groaning about how little time there is to learn everything!
🧶knitting/crochet/cross-stitch
🦻accessibility esp. in maths
👨🏫maths education
🧮operations research
🎵classical music
And am always groaning about how little time there is to learn everything!
Strategy "Don't go home until I've conquered this stupid mental hump that my brain is having trouble getting over, so that I could make headway on important work" was successful.
But at what cost...🥲
But at what cost...🥲
November 26, 2025 at 9:19 AM
Strategy "Don't go home until I've conquered this stupid mental hump that my brain is having trouble getting over, so that I could make headway on important work" was successful.
But at what cost...🥲
But at what cost...🥲
I realise I'm only putting to words now what other academics probably already know: I'm so frustrated at how hard it is to find people working on certain niche topics.
Can't tell if I'm looking in the wrong places or some people are just busy enough in their own bubbles.
Can't tell if I'm looking in the wrong places or some people are just busy enough in their own bubbles.
November 25, 2025 at 10:24 AM
I realise I'm only putting to words now what other academics probably already know: I'm so frustrated at how hard it is to find people working on certain niche topics.
Can't tell if I'm looking in the wrong places or some people are just busy enough in their own bubbles.
Can't tell if I'm looking in the wrong places or some people are just busy enough in their own bubbles.
Someone please tell me why ALL the opportunities are rolling in now when I'm like super busy and starting to feel the end-of-year burnout creeping in 😭
"When it rain, it pours" ringing true here, grrrr.
"When it rain, it pours" ringing true here, grrrr.
November 25, 2025 at 12:03 AM
Someone please tell me why ALL the opportunities are rolling in now when I'm like super busy and starting to feel the end-of-year burnout creeping in 😭
"When it rain, it pours" ringing true here, grrrr.
"When it rain, it pours" ringing true here, grrrr.
This was a good read. It's making me think about how to target other demographics (e.g. people with disabilities, in my case) for similar interventions.
How can we flip the decline in girls and gender-diverse students doing mathematics at higher levels?
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
doi.org
November 24, 2025 at 6:59 AM
This was a good read. It's making me think about how to target other demographics (e.g. people with disabilities, in my case) for similar interventions.
Reposted by Matt Mack
How can we flip the decline in girls and gender-diverse students doing mathematics at higher levels?
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
doi.org
November 24, 2025 at 4:25 AM
How can we flip the decline in girls and gender-diverse students doing mathematics at higher levels?
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
Zsuzsanna Dancso explains how her program of mentorship and encouragement piloted at
@sydney.edu.au increased enrolments from 22% to 30%📈 🧮
More here:
doi.org/10.64628/AA....
#MathSky
doi.
I understand why my work laptop has a program on it that forces me to take a break from it but there are moments where I do take a break but my brain has forgotten what I was doing so now it takes longer to resume the task arrrrrrrrrrrrgh /flails
November 24, 2025 at 1:30 AM
I understand why my work laptop has a program on it that forces me to take a break from it but there are moments where I do take a break but my brain has forgotten what I was doing so now it takes longer to resume the task arrrrrrrrrrrrgh /flails
It took me like two weeks from the start of seemingly random angry feelings I'd been having to finally having a breakthrough on realising where it was coming from. It was about 3-4 levels deep inside my psyche, ugh.
Taking tips on making emotional work easier.
Taking tips on making emotional work easier.
November 23, 2025 at 1:39 AM
It took me like two weeks from the start of seemingly random angry feelings I'd been having to finally having a breakthrough on realising where it was coming from. It was about 3-4 levels deep inside my psyche, ugh.
Taking tips on making emotional work easier.
Taking tips on making emotional work easier.
I'm wondering how this would ideally work. Digital signatures generated on one machine can be spoofed/generated-to-fit (eventually). If I wanted it to be more rock solid I can only imagine some cryptographic protocol requiring at least the receiving party to have their own private/public keys.
Image screening is going to fail. We need audit trails for data provenance.
I tried an even harder example on Gemini Pro image generation and this is quite scary/amazing. I asked for a microscopy image of around 20 HeLa cells, GFP tagged 20% nuclear, 10% membrane, +1 nuclear staining, + overlap. Image below and prompt in the following post.
November 22, 2025 at 10:29 PM
I'm wondering how this would ideally work. Digital signatures generated on one machine can be spoofed/generated-to-fit (eventually). If I wanted it to be more rock solid I can only imagine some cryptographic protocol requiring at least the receiving party to have their own private/public keys.
Reposted by Matt Mack
Image screening is going to fail. We need audit trails for data provenance.
I tried an even harder example on Gemini Pro image generation and this is quite scary/amazing. I asked for a microscopy image of around 20 HeLa cells, GFP tagged 20% nuclear, 10% membrane, +1 nuclear staining, + overlap. Image below and prompt in the following post.
November 22, 2025 at 4:21 PM
Image screening is going to fail. We need audit trails for data provenance.
Was looking for various theses to figure out the overall structure for mine. Randomly stumbled across a mathematical biology PhD thesis of someone who'd done 5 papers. Took me a while to recalibrate myself and acknowledge I probably won't be /that/ productive for my PhD 😅
November 22, 2025 at 9:43 AM
Was looking for various theses to figure out the overall structure for mine. Randomly stumbled across a mathematical biology PhD thesis of someone who'd done 5 papers. Took me a while to recalibrate myself and acknowledge I probably won't be /that/ productive for my PhD 😅
Reposted by Matt Mack
I think it’s something he and Mamdani have in common — they believe hard that leadership requires love for, joy in, and appreciation of the world as it is, because if you hate that for which you are responsible, you cannot see its future and hope to make it better. You will only destroy it.
November 22, 2025 at 12:18 AM
I think it’s something he and Mamdani have in common — they believe hard that leadership requires love for, joy in, and appreciation of the world as it is, because if you hate that for which you are responsible, you cannot see its future and hope to make it better. You will only destroy it.
As twitchy as I am about AI implementation in "everything" I can definitely get along with this thinking given that Firefox is open-source.
I think there's a more credible path to containing risk for users of AI if they do so on a browser created by Mozilla than on a browser created by OpenAI or Perplexity. And I don't think we're going to have zero users of AI. So that's why I think it's a good idea.
November 22, 2025 at 1:03 AM
As twitchy as I am about AI implementation in "everything" I can definitely get along with this thinking given that Firefox is open-source.
idk why I'm lining up this obviously-going-to-be-a-tearjerker-of-a-movie to watch tonight: "Big World" (2014), Chinese movie about a man with cerebral palsy.
If I don't post further tonight I'll probably be curled up in bed in emotional pain 🙃
www.imdb.com/title/t...
If I don't post further tonight I'll probably be curled up in bed in emotional pain 🙃
www.imdb.com/title/t...
November 21, 2025 at 9:39 AM
idk why I'm lining up this obviously-going-to-be-a-tearjerker-of-a-movie to watch tonight: "Big World" (2014), Chinese movie about a man with cerebral palsy.
If I don't post further tonight I'll probably be curled up in bed in emotional pain 🙃
www.imdb.com/title/t...
If I don't post further tonight I'll probably be curled up in bed in emotional pain 🙃
www.imdb.com/title/t...
Reposted by Matt Mack
Looking forward to next week as we kickstart the Biometrics conference!
We’ve also got an explicit printout of the Pac-Man rule to make it a welcome and inclusive event 🤗
#BIBC2025
We’ve also got an explicit printout of the Pac-Man rule to make it a welcome and inclusive event 🤗
#BIBC2025
November 21, 2025 at 5:10 AM
Looking forward to next week as we kickstart the Biometrics conference!
We’ve also got an explicit printout of the Pac-Man rule to make it a welcome and inclusive event 🤗
#BIBC2025
We’ve also got an explicit printout of the Pac-Man rule to make it a welcome and inclusive event 🤗
#BIBC2025
Some additional info on this for Australians: there's a National Building Code 2022 (yikes, so late?!) that dictates a standard, but in general if you can host conferences in newer buildings the more likely they'll have more accessibility considerations.
Can we please discuss accessibility at conferences?
When organizers and schedules assume that everyone can walk fast and climb stairs, people with mobility issues (visible or invisible) might feel excluded.
/
When organizers and schedules assume that everyone can walk fast and climb stairs, people with mobility issues (visible or invisible) might feel excluded.
/
November 20, 2025 at 9:24 AM
Some additional info on this for Australians: there's a National Building Code 2022 (yikes, so late?!) that dictates a standard, but in general if you can host conferences in newer buildings the more likely they'll have more accessibility considerations.
This morning I was helping a (neurodivergent) friend calm down from someone else spreading lies about him.
This afternoon my (also neurodivergent) housemate is doing his 6 month PhD confirmation talk, but I also know he's quietly freaking out about other things.
Hmm. 😵💫
This afternoon my (also neurodivergent) housemate is doing his 6 month PhD confirmation talk, but I also know he's quietly freaking out about other things.
Hmm. 😵💫
November 18, 2025 at 10:26 PM
This morning I was helping a (neurodivergent) friend calm down from someone else spreading lies about him.
This afternoon my (also neurodivergent) housemate is doing his 6 month PhD confirmation talk, but I also know he's quietly freaking out about other things.
Hmm. 😵💫
This afternoon my (also neurodivergent) housemate is doing his 6 month PhD confirmation talk, but I also know he's quietly freaking out about other things.
Hmm. 😵💫
Yesterday I learnt that mathematical (computational?) neuroscience builds off non-linear dynamical systems via this Reddit thread (which as of writing has two resource/expert recommendations), fascinating!
www.reddit.com/r/mat...
www.reddit.com/r/mat...
November 16, 2025 at 8:59 PM
Yesterday I learnt that mathematical (computational?) neuroscience builds off non-linear dynamical systems via this Reddit thread (which as of writing has two resource/expert recommendations), fascinating!
www.reddit.com/r/mat...
www.reddit.com/r/mat...
So if you don't know me, I'm hard-of-hearing and can only hear out of my left ear with a hearing aid, and for academic talks I use Google's Live Transcribe on my phone.
How does this impact where I choose to sit for academic talks? Here's a blog post!
matthras.com/2025/11...
How does this impact where I choose to sit for academic talks? Here's a blog post!
matthras.com/2025/11...
November 16, 2025 at 1:58 AM
So if you don't know me, I'm hard-of-hearing and can only hear out of my left ear with a hearing aid, and for academic talks I use Google's Live Transcribe on my phone.
How does this impact where I choose to sit for academic talks? Here's a blog post!
matthras.com/2025/11...
How does this impact where I choose to sit for academic talks? Here's a blog post!
matthras.com/2025/11...
If anyone can provide insight on how diVine can ban/prevent AI-generated videos on a technical/code/metadata level I'd love to know because I don't know enough on that front. Any solution I can think of is easily circumvented/bypassed.
November 15, 2025 at 1:48 AM
If anyone can provide insight on how diVine can ban/prevent AI-generated videos on a technical/code/metadata level I'd love to know because I don't know enough on that front. Any solution I can think of is easily circumvented/bypassed.
This week's moment that made me feel old: Watching "Man vs. Bee" and realising there're kids today who haven't seen the original Mr. Bean nor the Mr. Bean movies.
I do wonder if I'd enjoy it more if I didn't recognise all the recycled bits 🤔
I do wonder if I'd enjoy it more if I didn't recognise all the recycled bits 🤔
November 13, 2025 at 11:59 PM
This week's moment that made me feel old: Watching "Man vs. Bee" and realising there're kids today who haven't seen the original Mr. Bean nor the Mr. Bean movies.
I do wonder if I'd enjoy it more if I didn't recognise all the recycled bits 🤔
I do wonder if I'd enjoy it more if I didn't recognise all the recycled bits 🤔
I'm noticing that discussions at a high/policy/leadership level using generalised terms run into disagreements purely because different people work off different assumptions underlying said generalised terms and it's really annoying me 💢
1/2
1/2
November 12, 2025 at 10:42 PM
I'm noticing that discussions at a high/policy/leadership level using generalised terms run into disagreements purely because different people work off different assumptions underlying said generalised terms and it's really annoying me 💢
1/2
1/2
Reposted by Matt Mack
UK Maths students - interested in doing a cross-disciplinary PhD in biofilms and doing part of your studies in Australia? This could be your chance!
ima.org.uk/27939/compet...
ima.org.uk/27939/compet...
Competition funded PhD position in Mathematical modelling of biofilms
ima.org.uk
November 11, 2025 at 10:47 PM
UK Maths students - interested in doing a cross-disciplinary PhD in biofilms and doing part of your studies in Australia? This could be your chance!
ima.org.uk/27939/compet...
ima.org.uk/27939/compet...
First time coming into work and seeing two fire engines right outside my building 😳.
Thankfully the building affected was the one on the opposite side of the walk way 😅
That'll be one heck of a hump day for the week!
Thankfully the building affected was the one on the opposite side of the walk way 😅
That'll be one heck of a hump day for the week!
November 11, 2025 at 11:24 PM
First time coming into work and seeing two fire engines right outside my building 😳.
Thankfully the building affected was the one on the opposite side of the walk way 😅
That'll be one heck of a hump day for the week!
Thankfully the building affected was the one on the opposite side of the walk way 😅
That'll be one heck of a hump day for the week!
I'm really appreciating that I'm learning data cleaning + stats + machine learning stuff through my PhD by analysing clinical data but it's also really affirming my "I really couldn't be a generic data scientist unless the data is relevant to something I care about" 😅
November 11, 2025 at 10:07 AM
I'm really appreciating that I'm learning data cleaning + stats + machine learning stuff through my PhD by analysing clinical data but it's also really affirming my "I really couldn't be a generic data scientist unless the data is relevant to something I care about" 😅
Interesting paper: Researchers wondered if it's possible to predict "Eureka!" moments based on observable behaviour (on a blackboard), so they recorded mathematicians working on Putnam problems. tl;dr: Definitely some hints!
www.pnas.org/doi/10....
www.pnas.org/doi/10....
November 10, 2025 at 9:52 PM
Interesting paper: Researchers wondered if it's possible to predict "Eureka!" moments based on observable behaviour (on a blackboard), so they recorded mathematicians working on Putnam problems. tl;dr: Definitely some hints!
www.pnas.org/doi/10....
www.pnas.org/doi/10....