if new yorkers really don't want to be americans, they should organize a vote to secede and ask to join canada. i'd be happy to admit new york as a province of canada, perhaps combined with a few other states, if they would like to come, too.

https://ca.news.yahoo.com/canadians-don-t-want-come-090023329.html

it looks like avi lewis is going to win the leadership. but, i would expect that heather macpherson will keep her seat in the legislature in the next election, and avi lewis will not succeed in winning one for himself.

the ndp are consequently now faced with a problem that you see in canada with what we call fringe parties, which is that they can get an mp elected here and there, but they can't get their leader elected. we saw this most recently with the green party, and they had to go back to elizabeth may.

mr. lewis has already made attempts to win seats twice, in 2021 and 2025, and lost both times. there is no obvious place for mr. lewis to run where he might win. i frankly don't think there is a riding in the country that he's likely to win in.

further, mr. lewis' ethnicity (i don't suspect he's very religious) is likely to generate a large amount of hostility amongst core ndp voters, and i'd get some popcorn for that, as it's likely to be entertaining watching canada's previously left-wing party collapse into far-right conspiracy theory rhetoric. but it's coming.

i don't exactly disagree with the outcome. mr lewis was clearly the best candidate. but he's also the most unelectable candidate in canada in 2026 that you could imagine creating out of clay, if you could, and it speaks to a contradiction in the ndp voting coalition, and donor base. lewis will be able to hold 10-15% of voters across the country, but the ridings the ndp has been able to hold since mulcair have tended to be a little right of centre and more traditionally conservative, and the mps it actually elects reflect that. it follows that the mps that the ndp is able to elect are fairly centrist, while it's donors are more to the left, and there doesn't seem to be a way to address that other than to pick a fight with the liberal party, which the ndp tends to lose when it does, because the liberals constantly run to the left of the ndp on issues that actually matter, and the liberals have a far more left-wing record in actual governance. it's hard for an informed leftist in canada to actually vote ndp, given their record in government. the ndp actually have this really bad habit of unexpectedly swinging hard to austerity. at this point, they can't blame it on bad apples. it's happened across the country, and it leaves them in ruins. 

a lot of former ndp voters are now conservative voters, and avi lewis is not going to swing them back, so how does he win a seat? he can't. and he won't.

this is the deep secret of canadian politics: the ndp is not a socialist party but is actually a real conservative party. not a post-thatcherian neo-liberal party, a pre-clarkian, stanfeldian, diefenbakerian old-timey conservative party. 

so, they're kind of fucked.

lewis wants to be the guy that saves the ndp, but he's going to be the guy that buries them, and what happens next to the ndp is going to be ugly. he's going to win the vote, and then get attacked by his own party, which doesn't want him to win, and it's going to get vicious. this is going to alienate his donors, who are going to bolt to the greens or liberals (or bloc) and it's going to leave behind the ugliest part of the ndp, to collapse into a new socred party, and then fade into obscurity.

there may be some more floor crossings from the ndp to the liberals. unfortunately. because canadians didn't vote for a majority, but that will make it real, until the next election.

trump didn't underestimate iran, he overestimated them - he decided they were rational and would behave civilly, and they aren't and haven't.

i admit i never got over the death of optimus prime. 

i was once so full of hope and joy. it shattered me, and i've never been the same since.

it is the source of my deeply cynical, existentialist outlook on life and the reason i can't take anything seriously. it taught me that nothing matters.

i cried my little eyes out.

with the exception of a short time window from about 1993 to about 1997, about when norm got cancelled, and when i was in my early teens, i actually know almost nothing about saturday night live.

it was on saturday nights.

i didn't tend to find myself in front of a tv at midnight on sunday morning. 

we didn't have youtube back then, and i didn't have a pvr. i tuned out.

i probably would have watched it more if it was on sunday night or a night in the week. i watched a lot of conan when i was a kid, because it was on during the week.

in hindsight, that was kind of a high point for the show, i believe. they kind of lost the plot when they axed norm. and people started dying, which makes for bad comedy. i do have an opinion on cast members like will farrell and jimmy fallon, and it's not remotely positive. these were not funny people. at all.

but i just didn't watch it after about 1998, and nobody i knew watched it, and it never got spoken of. i really have nothing to say about it.

i have seen more reruns of sctv or monty python or older snl than i have of post-norm snl.

most of these colleges are diploma mills and should close. the provincial governments should step in and revoke their licenses. if you can't exist without importing D students from india, you shouldn't exist at all. these are students that can't get in to real schools.

on the other hand, the real schools that will survive this should be taking the down time as an opportunity to build student housing, to prevent low income canadian citizens from being forced to compete for rental housing with temporary residents.

if you need to import students to prevent your third rate school from going bankrupt, house them on your campus, and don't force rent inflation on canadian citizens. that's not a valid economic model.

further, third rate students from third world countries are not an acceptable source of cheap labour for small businesses. these small businesses need to follow proper labour practices and hire canadians at legal wages. if these small businesses can't pay legal wages, they should close.

small businesses are bad for labour rights because they're not unionized. we should be slowly getting rid of small businesses in favour of conglomerates with unionized workforces, not propping them up by importing slaves for them to exploit. if your business model needs slaves to be profitable, your business shouldn't exist.

https://www.ctvnews.ca/canada/article/canada-as-a-brand-has-suffered-post-secondary-schools-react-to-ag-report-on-international-students/

no. bill 18 is a dangerous attack on the rights of disabled people to self-determination, and there is a broad consistency developing - the canadian right, which includes the liberal party, consistently refuses to accept the concept of individual rights, both at home and abroad, and is instead consistently enforcing religious values about "dignity" on people that have won the constitutional right to be free from these religious beliefs.

the federal government needs to stand up for the rights of the disabled to kill themselves on demand like any other group, and take the albertan government to court to force them to enforce the law.

https://calgaryherald.com/opinion/columnists/pro-bill-18-alberta-moves-to-limit-assisted-death-now-canada-must-follow

i wish i had used the benzalkonium chloride earlier. it cleared it right up, in a few spots. i'll keep at it until it's gone.

it doesn't tell me what it was, as it would kill anything, but it tells me it probably wasn't the ants, bed bugs, fleas or anything else like that.

i'm glad that's dealt with as it was freaking me out.

if you want better error handling, you hire C programmers.

if there are any still alive.

there are people thinking about this theoretically, which is good. i mean, who is writing these ai programs? kids with degrees in oracle or java? they won't get this.

i would assume this is going to get fixed.

the companies need to come clean on this.

The impossibility of having a general method that can assert whether an arbitrary AI is aligned or not does not mean that it is impossible to construct an AI that is provably aligned. Instead, it should be interpreted that there exist many AIs that cannot be proven to be aligned or not, while there is also a countable set of AIs that are proven to be aligned. Therefore, it is our objective to develop and utilize such a countable set of proven aligned AIs. The architecture and its development process are fundamental to ensuring safety.

Developing an AI model that always halts allows for the alignment and other properties of the AI model to be asserted computationally, a task that would be computationally impossible for arbitrary models.

https://www.nature.com/articles/s41598-025-99060-2

yeah, i've looked into this idea of "ai scheming" and it's nowhere near as complex or profound as some people want to decide it is. i think people are looking for sentience and finding it where it doesn't exist. it says more about the analysts. 

the good news is that it suggests incompetence rather than malice. these are not demonstrations of sentience. they're essentially programming errors.

in every example presented, one of two things are true:

1) the ai was programmed to lie, and it did.

2) the ai was asked to solve a problem that computer scientists call undecidable and it creates an error that is being misinterpreted as the ai being dishonest, because it's what the researchers want to see.

what is an undecidable problem?

In computability theory and computational complexity theory, an undecidable problem is a decision problem for which it is proved to be impossible to construct an algorithm that always leads to a correct yes-or-no answer. 

this may be a mindfuck to people, because computers can solve everything. right? no. in fact, computers can't even do basic math because they're base 2. computers are constantly making simple mistakes, and we're forced to write incredibly complicated software, and design elaborate pieces of hardware, to catch all of the mistakes they make. end users don't see that, until they do. but computer bugs are so 90s.

when was the last time you saw a programming mistake make it to production?

well, that's exactly what "ai scheming" is.

the program should be doing better error correction in trying to catch these undecidable problems and guiding them to determined outcomes, or in preventing users from breaking the algorithm. these are programming mistakes. the software engineers should not be tripping out on them, they should be correcting them.

now, computability theory is a pretty big branch of applied mathematics and it's pretty established and in fact pretty old. one of the results of computability theory is that there are uncountably many undecidable problems. so you can't catch every mistake. 

but you can design the program to tell the user that it broke it. and i think doing that would be extremely helpful in training users to interact with the program correctly.

not every request or command has an answer. if the program can't answer, it should say that. if the problem is undecidable, or creates a contradiction, it should say that. it shouldn't always produce a response.

not every question has an answer, and that is a fundamental result of modern mathematics. users should be trained by the ai to understand that.