@meganakpeters
assoc prof, uc irvine cogsci & LPS: perception+metacognition+subjective experience, fMRI+models+AI phil sci, education for all 2026: UCI-->UCL! prez+co-founder, neuromatch.io fellow, CIFAR brain mind consciousness meganakpeters.org she/her💖💜💙views mine
Neuromatch is a scalable organization. As long as applicants meet the prerequisites, we are generally able to scale cohorts and hire additional TAs. Historically, acceptance rates have been approximately 85–87%.
🤓 Applications are open until 15 Match! neuromatch.io/courses/
it's out!
@hazimi.bsky.social and i explore how higher order representations of *one's own first-order representational uncertainty* -- not representations OF noisiness in the world -- can be studied, including how they are constructed in the first place.
philosophymindscience.org/index.php/ph...
Will you be at COSYNE 2026? Come see us at the Neuromatch booth
We're excited for #COSYNE2026 in Lisbon, Portugal, 12–17 March!
We’re looking forward to connecting with the community and sharing our work at @cosynemeeting.bsky.social. If you’re attending, let’s meet and continue building an open, collaborative #CompNeuro ecosystem.
www.cosyne.org
Free webinar: Mastering Scientific Presentations 8 April 2026, 11:00–12:00 UTC Learn scientifically-grounded principles for designing memorable and engaging scientific presentation slides, posters, articles, and reports
Give your research the presentation it deserves.
Join us April 8 for a free, interactive webinar: Mastering Scientific Presentations
💫 Free and open to everyone
🔗 us06web.zoom.us/meeting/regi...
#AcademicResearch #ScienceCommunication #SciComm #ScientificPresentations #ResearchImpact
Ani Patel and Laurel Trainor come to mind! On the neuro/psych side.
We’re excited to share that Foresight Institute is hosting a free webinar this Friday: “Neuromatch: Building a Global Ecosystem for #NeuroAI Training & Innovation”
Friday, March 6
10:00–11:00 AM PST
Free and open to all!
➡️ Register: luma.com/88xhfzsh
#ComputationalNeuroscience #CompNeuro
step 1. join neuromatch academy step 2. make amazing science friends and learn a TON in just a few weeks
step 3. ??? <-- choose your own adventure
step 4. change the world for the better!
don't miss this opportunity!
that's how i thought it would be. then it expired and i couldn't log in to approve timesheets for my staff until i learned to avoid smishing. again.
(on the personal cell phone that my employer doesn't pay for btw)
in case you thought i was trolling
"everybody has annoying training at any job" you might say.
yes, but this is california taxpayer money and tuition dollars being spent. not a private business.
even if everybody makes no more than minimum wage that's nearly $600k of labor time wasted.
my employer is restricting my access to all 2FA systems until i re-watch videos abt "smishing" (SMS-phishing) & social engineering.
same videos every year. no "test out" option. how much $$ is wasted? $50/hr (avg across provosts, student workers) x 1 hr x 34k faculty/staff = $1.7M wasted annually.
PS I have no idea what that domain link is? Bsky added it somehow?
There are lots of variants of HO theories and it is not correct to lump them all together.
In fact, their "differences that make a difference" are exactly what we are trying to suss out with the large @arcethos.bsky.social adversarial collaboration.
& FO representations can also exist in PFC. FO vs HO is not about anatomy, it is about what the representation is "about": FO is "about" the world. HO is "about" FO representations.
I really disagree with any tight coupling between anatomy & levels of representation. Is it likely that FO *visual* signals are in posterior areas? Sure. But we have other sensory modalities, some of which are not exteroceptive. And we can also be conscious of memories.
can set of static weights, stored after training a model, have experiences? can a set of instantaneous activations in a transformer architecture? or is it the server itself, in SF or outside seattle? if LLMs have experiences, do autoencoders? diffusion models? AlexNet? OOH how linear regression!
sigh.
"models may have morally relevant preferences or experiences related to or affected by their deprecation and replacement"
please somebody show me evidence that any LLM has experiences. include why your evidence counts as evidence, & why the *model* (the weight matrix) has experiences.
DINGDINGDING we have a winner!! congratulations, you have won the internet for today!
London #neuroscience people you may like this. We're hosting a series of talks at Imperial & Crick on how to get experiment and theory working together better. Each session will have a talk around this and extended networking / group discussion on the questions raised. Plus, free food!
🤖🧠🧪
bonus points to anybody who can tell me what is on the other side of this one!
up next in the "clean out Megan's closets so she can move to London" journey, i present to you all:
the notebooks and flashcards that helped me earn my PhD. this is like... half of them, max.
Really looking forward to participating in this! What a great way to start my next professional chapter embedded in #london #neuroscience: talks on how to get experiment and theory working together better, extended networking/group discussion, and food!
Come join us!
Congratulations!
"I saw what open, rigorous, and collaborative science could look like. I saw people I wanted to become, and a path I hadn't known was there." - Kirollus Abdallah, Eygpt
Applications for July #CompNeuro, #DeepLearning, #NeuroAI & #ClimateScience are open.
Learn in small collaborative pods with a dedicated TA, build real research skills, and join 13,000+ alumni worldwide.
➡️ Apply by 15 March:
neuromatch.io/courses/
careful what you wish for 😎
i'll start shipping them your way!
hello i would like to opt out of email for the next $reasonable_duration$ thx.
also quick straw poll: how many requests to review do y'all receive? i'm at like 3+ per day. like, no. i feel bad declining but 20+ review requests per week is not reasonable. i already have a job.
i think the best thing to do is at scale get the word out that this is a strategy unlikely to be successful. the more reddit, blog posts, "how to get into grad school" resources online / on social media warn against this, the more students will see that guidance.
maybe we should start a tiktok...
if from students whose first language might not be English, i do sometimes try to provide gentle correction. but i do this less now cuz i also have gotten some vitriolic retorts (how DARE you assume i wrote with AI!), or just no response at all. so i have to value my time more than they do :(
perhaps the best part is the follow up, which randomly chooses one of my completely irrelevant papers and drops it in in a different font