joel
@jmoonz.bsky.social
There is also a torrent at magnet:?xt=urn:btih:8c6aa712c99d1240bc116447d6ec67d08cb16913&xt=urn:btmh:1220b5e28de225fe8035e9ec46fb41adb3a4898680c6b0085e5895e911f8ef4000ab&dn=Epstein%20Estate%20Documents%20-%20Seventh%20Production
Since it can be annoying to download manually.
Since it can be annoying to download manually.
November 12, 2025 at 11:35 PM
There is also a torrent at magnet:?xt=urn:btih:8c6aa712c99d1240bc116447d6ec67d08cb16913&xt=urn:btmh:1220b5e28de225fe8035e9ec46fb41adb3a4898680c6b0085e5895e911f8ef4000ab&dn=Epstein%20Estate%20Documents%20-%20Seventh%20Production
Since it can be annoying to download manually.
Since it can be annoying to download manually.
plus you can also quit if they become annoying.
November 12, 2025 at 10:53 PM
plus you can also quit if they become annoying.
Oh, and sensitivity analysis is the other big one. If your code is off, sensitivity analysis can just magnify any code error... or when your results don't change at all, point to a bigger issue of the code not doing what you thought.
November 12, 2025 at 7:23 PM
Oh, and sensitivity analysis is the other big one. If your code is off, sensitivity analysis can just magnify any code error... or when your results don't change at all, point to a bigger issue of the code not doing what you thought.
Which is understandable, many PhDs use coding as a tool and have not been taught "good" software structure. They have many other advanced topics to try to get their head around!
Code that works and gives a reasonable answer is often good enough.
Code that works and gives a reasonable answer is often good enough.
November 12, 2025 at 7:21 PM
Which is understandable, many PhDs use coding as a tool and have not been taught "good" software structure. They have many other advanced topics to try to get their head around!
Code that works and gives a reasonable answer is often good enough.
Code that works and gives a reasonable answer is often good enough.
Yeah exactly - although I learnt during my PhD that even research code is very rarely write once.
Replication, retries, bug fixes, meant it was worth spending a bit of time making it clear and understandable, with tests.
Even if just for me to have confidence in my paper claims and results!
Replication, retries, bug fixes, meant it was worth spending a bit of time making it clear and understandable, with tests.
Even if just for me to have confidence in my paper claims and results!
November 12, 2025 at 7:18 PM
Yeah exactly - although I learnt during my PhD that even research code is very rarely write once.
Replication, retries, bug fixes, meant it was worth spending a bit of time making it clear and understandable, with tests.
Even if just for me to have confidence in my paper claims and results!
Replication, retries, bug fixes, meant it was worth spending a bit of time making it clear and understandable, with tests.
Even if just for me to have confidence in my paper claims and results!
As a PhD, who has also later had to build on other PhD/researcher code, I would suggest it's the style of the average PhD.
Good PhD code is the exception.
Good PhD code is the exception.
November 12, 2025 at 7:04 PM
As a PhD, who has also later had to build on other PhD/researcher code, I would suggest it's the style of the average PhD.
Good PhD code is the exception.
Good PhD code is the exception.
Please release in New Zealand - Gabe seemed to enjoy his time here during Covid and it's sad to always miss out on official distribution of Steam hardware.
November 12, 2025 at 6:13 PM
Please release in New Zealand - Gabe seemed to enjoy his time here during Covid and it's sad to always miss out on official distribution of Steam hardware.
Yeah, I felt like everything was going janky before LLMs, but it's definitely accelerated the jank
November 12, 2025 at 5:32 PM
Yeah, I felt like everything was going janky before LLMs, but it's definitely accelerated the jank
champion, will download soon and seed
November 12, 2025 at 5:31 PM
champion, will download soon and seed
If that happens I'm kinda bullish for AI advancement overall.
We'll get a flood of creativity when people can run these in their homelabs without remortgaging their house.
...I can dream about it anyway.
We'll get a flood of creativity when people can run these in their homelabs without remortgaging their house.
...I can dream about it anyway.
November 12, 2025 at 5:09 AM
If that happens I'm kinda bullish for AI advancement overall.
We'll get a flood of creativity when people can run these in their homelabs without remortgaging their house.
...I can dream about it anyway.
We'll get a flood of creativity when people can run these in their homelabs without remortgaging their house.
...I can dream about it anyway.
Unless entirely freeze dried, then that's a reasonable ratio
November 12, 2025 at 4:31 AM
Unless entirely freeze dried, then that's a reasonable ratio
So many calls lately for US government to save the bubble.
November 11, 2025 at 5:41 PM
So many calls lately for US government to save the bubble.
It requires post separate from comments. Up/Down voting. And threaded comments.
While many social networks had these, almost all of them set these on fire making their networks less functional but higher "engagement".
While many social networks had these, almost all of them set these on fire making their networks less functional but higher "engagement".
November 10, 2025 at 6:53 PM
It requires post separate from comments. Up/Down voting. And threaded comments.
While many social networks had these, almost all of them set these on fire making their networks less functional but higher "engagement".
While many social networks had these, almost all of them set these on fire making their networks less functional but higher "engagement".
Normal linear algebra is easier than heteroscedastic linear algebra.
Ba-dum-tsch
Ba-dum-tsch
November 8, 2025 at 7:35 AM
Normal linear algebra is easier than heteroscedastic linear algebra.
Ba-dum-tsch
Ba-dum-tsch
ah right, I understand now.
November 7, 2025 at 11:14 PM
ah right, I understand now.
I don't know what you're talking about. Who's talking about retrospectively taking people's lifetime income?
November 7, 2025 at 10:46 PM
I don't know what you're talking about. Who's talking about retrospectively taking people's lifetime income?
Nah - it'd be a pretty sweet ride to get a salary for the rest of your life for doing a single term in government.
November 7, 2025 at 10:45 PM
Nah - it'd be a pretty sweet ride to get a salary for the rest of your life for doing a single term in government.
Yeah I jumped the gun sorry.
Still - they inevitable will off load this under some new instrument name.
Still - they inevitable will off load this under some new instrument name.
November 7, 2025 at 7:57 PM
Yeah I jumped the gun sorry.
Still - they inevitable will off load this under some new instrument name.
Still - they inevitable will off load this under some new instrument name.
Investment bankers - and banking generally - have learnt debt is a great way to make money, especially if you can off load the majority of risk to others. Ideally while masking the real risk with complex intruments so that you can earn a premium.
November 7, 2025 at 7:53 PM
Investment bankers - and banking generally - have learnt debt is a great way to make money, especially if you can off load the majority of risk to others. Ideally while masking the real risk with complex intruments so that you can earn a premium.
Well tracker networks do the same with differential pricing.
Can you afford to pay more for a holiday? Then you most certainly will!
But it's not about the identity of the person, it's about "how can we maximise profit, be judging price vs conversion rate for this tranche of customer"
Can you afford to pay more for a holiday? Then you most certainly will!
But it's not about the identity of the person, it's about "how can we maximise profit, be judging price vs conversion rate for this tranche of customer"
November 7, 2025 at 7:45 PM
Well tracker networks do the same with differential pricing.
Can you afford to pay more for a holiday? Then you most certainly will!
But it's not about the identity of the person, it's about "how can we maximise profit, be judging price vs conversion rate for this tranche of customer"
Can you afford to pay more for a holiday? Then you most certainly will!
But it's not about the identity of the person, it's about "how can we maximise profit, be judging price vs conversion rate for this tranche of customer"
That's generally what an effective tax system tracks.
Of course, you can't control people who work outside of the law and refuse to report their earnings.
Of course, you can't control people who work outside of the law and refuse to report their earnings.
November 7, 2025 at 7:39 PM
That's generally what an effective tax system tracks.
Of course, you can't control people who work outside of the law and refuse to report their earnings.
Of course, you can't control people who work outside of the law and refuse to report their earnings.
I think that salary should be linked to the average income of the citizens.
It can be by a multiplier if needed, but need to encourage long term and society wide planning.
It can be by a multiplier if needed, but need to encourage long term and society wide planning.
November 7, 2025 at 5:58 PM
I think that salary should be linked to the average income of the citizens.
It can be by a multiplier if needed, but need to encourage long term and society wide planning.
It can be by a multiplier if needed, but need to encourage long term and society wide planning.