Arun Niranjan
@arun-niranjan.bsky.social
Software engineer and former neuroscientist
Also if you get to Osaka and it isn't your vibe, I highly recommend checking Kobe out (it was my favourite city this trip)
November 25, 2025 at 1:09 PM
Also if you get to Osaka and it isn't your vibe, I highly recommend checking Kobe out (it was my favourite city this trip)
November 25, 2025 at 1:07 PM
Ah these were the major cities I didn't spend any time in this trip! If you are in shibuya in Tokyo and haven't checked it out, this izakaya is great
maps.app.goo.gl/UDvFDigMYcSA...
maps.app.goo.gl/UDvFDigMYcSA...
maps.app.goo.gl
November 25, 2025 at 1:06 PM
Ah these were the major cities I didn't spend any time in this trip! If you are in shibuya in Tokyo and haven't checked it out, this izakaya is great
maps.app.goo.gl/UDvFDigMYcSA...
maps.app.goo.gl/UDvFDigMYcSA...
What cities are you in? Just Tokyo or further afield?
November 25, 2025 at 12:28 PM
What cities are you in? Just Tokyo or further afield?
Are there any other countries which apply these sorts of controls to foreign nationals?
November 25, 2025 at 11:16 AM
Are there any other countries which apply these sorts of controls to foreign nationals?
No rush, I have my ugly python script in a git branch, I can resurrect it whenever
November 25, 2025 at 11:09 AM
No rush, I have my ugly python script in a git branch, I can resurrect it whenever
I just got back from Japan and my wife is vegetarian - lmk if you need recommendations 🙏🏾
November 25, 2025 at 11:01 AM
I just got back from Japan and my wife is vegetarian - lmk if you need recommendations 🙏🏾
Where would you rank *yourself* playing Kotor 2?
November 24, 2025 at 10:14 PM
Where would you rank *yourself* playing Kotor 2?
Reposted by Arun Niranjan
As my colleague Chris Giles put it: "The economics of this is pretty simple. 1. You should be a billionaire if you're worth it 2. Probably, no one is 3. If you're not worth it and a billionaire - you've got your money from rent exploitation - which should be stopped/taxed"
November 23, 2025 at 11:28 PM
As my colleague Chris Giles put it: "The economics of this is pretty simple. 1. You should be a billionaire if you're worth it 2. Probably, no one is 3. If you're not worth it and a billionaire - you've got your money from rent exploitation - which should be stopped/taxed"
But you can infer hfia from a questionnaire output, and easily extend the model to:
- combine with other modalities and hidden states
- decide on transition probabilities and interventions to try and *change* a hfia score
- relax the logical rules determining the score and learn them from data
- combine with other modalities and hidden states
- decide on transition probabilities and interventions to try and *change* a hfia score
- relax the logical rules determining the score and learn them from data
November 23, 2025 at 3:14 PM
But you can infer hfia from a questionnaire output, and easily extend the model to:
- combine with other modalities and hidden states
- decide on transition probabilities and interventions to try and *change* a hfia score
- relax the logical rules determining the score and learn them from data
- combine with other modalities and hidden states
- decide on transition probabilities and interventions to try and *change* a hfia score
- relax the logical rules determining the score and learn them from data
Anyhow, I managed to construct a basic pomdp with 9 observation modalities (1 for each frequency question) and 1 hidden state factor (the hfia score) and encode the logic rules in the state-observation likelihood matrix.
A glorified way of doing if statements basically, not hugely useful 😅
A glorified way of doing if statements basically, not hugely useful 😅
November 23, 2025 at 3:10 PM
Anyhow, I managed to construct a basic pomdp with 9 observation modalities (1 for each frequency question) and 1 hidden state factor (the hfia score) and encode the logic rules in the state-observation likelihood matrix.
A glorified way of doing if statements basically, not hugely useful 😅
A glorified way of doing if statements basically, not hugely useful 😅
And is the real question here how to make hfia an ordinal variable rather than categorical, or do we explicitly need to keep the underlying questionnaire observations that logically determine hfia?
November 23, 2025 at 10:11 AM
And is the real question here how to make hfia an ordinal variable rather than categorical, or do we explicitly need to keep the underlying questionnaire observations that logically determine hfia?
Having another think about this now I am home.
What's your overall goal? Do you have a bunch of survey data that you're trying to fit to an existing model?
What's your overall goal? Do you have a bunch of survey data that you're trying to fit to an existing model?
November 23, 2025 at 10:09 AM
Having another think about this now I am home.
What's your overall goal? Do you have a bunch of survey data that you're trying to fit to an existing model?
What's your overall goal? Do you have a bunch of survey data that you're trying to fit to an existing model?
And possibly only a strictly male advice to other men thing 🤷🏾♂️
The idea of believing that a pick up line would actually succeed feels icky to me
The idea of believing that a pick up line would actually succeed feels icky to me
November 21, 2025 at 10:56 PM
And possibly only a strictly male advice to other men thing 🤷🏾♂️
The idea of believing that a pick up line would actually succeed feels icky to me
The idea of believing that a pick up line would actually succeed feels icky to me
Careful what you wish for, you might end up with two week Agile sprints 😱
November 20, 2025 at 10:42 PM
Careful what you wish for, you might end up with two week Agile sprints 😱
Ah ignore me, turns out cloud flare doesn't like anyone's website 😅
November 18, 2025 at 2:00 PM
Ah ignore me, turns out cloud flare doesn't like anyone's website 😅
Looks like cloud flare doesn't like your site 😞
November 18, 2025 at 1:59 PM
Looks like cloud flare doesn't like your site 😞
The clif bar has simply been subsumed within the agents markov blanket
November 18, 2025 at 10:23 AM
The clif bar has simply been subsumed within the agents markov blanket
This must be this "embodied intelligence" concept I keep hearing about.
November 18, 2025 at 9:58 AM
This must be this "embodied intelligence" concept I keep hearing about.
Reposted by Arun Niranjan
"It is terrible that Shabana Mahmood faces overt racism for being a high profile Asian woman" = good argument
"The key to ending overt racism against Asians in Britain is that we must Stop the Boats" = a very bad argument indeed, on so many levels
"The key to ending overt racism against Asians in Britain is that we must Stop the Boats" = a very bad argument indeed, on so many levels
November 17, 2025 at 11:23 PM
"It is terrible that Shabana Mahmood faces overt racism for being a high profile Asian woman" = good argument
"The key to ending overt racism against Asians in Britain is that we must Stop the Boats" = a very bad argument indeed, on so many levels
"The key to ending overt racism against Asians in Britain is that we must Stop the Boats" = a very bad argument indeed, on so many levels