Bijan Parsia
@bparsia.bsky.social
I am a Prof of Comp Sci at Uni of Manchester, UK. UCU NEC disabled member rep (2023-24). UCUCommons. All posts are my own behalf alone. He/him/they.
Was it a good use of $20k? Marketing wise, maybe? Understandingwise, maybe not?
February 10, 2026 at 8:15 AM
Was it a good use of $20k? Marketing wise, maybe? Understandingwise, maybe not?
This is especially true when a company is selling something and hard.
I’m not going to say the compiler is crap or unworkable or unimpressive…I just don’t know and I don’t think we’re collectively in a good spot to know.
I’m not going to say the compiler is crap or unworkable or unimpressive…I just don’t know and I don’t think we’re collectively in a good spot to know.
February 10, 2026 at 8:15 AM
This is especially true when a company is selling something and hard.
I’m not going to say the compiler is crap or unworkable or unimpressive…I just don’t know and I don’t think we’re collectively in a good spot to know.
I’m not going to say the compiler is crap or unworkable or unimpressive…I just don’t know and I don’t think we’re collectively in a good spot to know.
In the early days of computer science you could get a paper (and sometimes a dissertation) for “Gosh, I wrote a program and it ran!” (slight exaggeration) but I don’t think it’s helpful for LLM evaluation.
February 10, 2026 at 8:15 AM
In the early days of computer science you could get a paper (and sometimes a dissertation) for “Gosh, I wrote a program and it ran!” (slight exaggeration) but I don’t think it’s helpful for LLM evaluation.
I saw that someone was able to compile SQLLite with it and get a working super duper slow binary out of it, which is good. I also saw people say it failed to compile a hello world, which maybe not so good.
February 10, 2026 at 8:15 AM
I saw that someone was able to compile SQLLite with it and get a working super duper slow binary out of it, which is good. I also saw people say it failed to compile a hello world, which maybe not so good.
Issues: No pre-registration of goals and success criteria. Many people have said “compile the Linux Kernal” doesn’t mean “you can use it to build the Linux kernel” or even “it thus will compile a lot of stuff”.
This makes things tricky from the outset.
This makes things tricky from the outset.
February 10, 2026 at 8:15 AM
Issues: No pre-registration of goals and success criteria. Many people have said “compile the Linux Kernal” doesn’t mean “you can use it to build the Linux kernel” or even “it thus will compile a lot of stuff”.
This makes things tricky from the outset.
This makes things tricky from the outset.
Here’s my comment on the Claude C compiler…thing.
It wasn’t a good experiment and thus is hard to evaluate for bad reasons. (It might also be hard to evaluate for good ie domain reasons.)
It wasn’t a good experiment and thus is hard to evaluate for bad reasons. (It might also be hard to evaluate for good ie domain reasons.)
February 10, 2026 at 8:15 AM
Here’s my comment on the Claude C compiler…thing.
It wasn’t a good experiment and thus is hard to evaluate for bad reasons. (It might also be hard to evaluate for good ie domain reasons.)
It wasn’t a good experiment and thus is hard to evaluate for bad reasons. (It might also be hard to evaluate for good ie domain reasons.)
Reposted by Bijan Parsia
A key thing about “mass deportation” is that it doesn’t work so you go from mass deportation to mass imprisonment without cause and sometimes to other more final solutions
Madagascar again? Or were they planning on a different destination this time?
July 18, 2024 at 1:18 PM
A key thing about “mass deportation” is that it doesn’t work so you go from mass deportation to mass imprisonment without cause and sometimes to other more final solutions
Reposted by Bijan Parsia
Government to cover 90% of existing council SEND debts | TES
Government to cover 90% of existing council SEND debts
Councils will be able to apply for a ‘High Needs Stability Grant’ if they get DfE approval for a SEND reform plan
www.tes.com
February 10, 2026 at 6:39 AM
Government to cover 90% of existing council SEND debts | TES
This is really bad.
1 in 10 students sees peers make sexual deepfakes | TES
1 in 10 students sees peers make sexual deepfakes
And a third of teenagers are using AI for mental health support, according to research into young people’s use of technology
www.tes.com
February 10, 2026 at 6:42 AM
This is really bad.
Reposted by Bijan Parsia
one thing that will always be a mystery to me is the number of trump voters who basically thought they were electing george bush again reut.rs/3Mc4h4H
As Trump presidency enters second year, his voters share hopes – and concerns
Trump voters urge him to tone down the rhetoric and focus more on domestic issues, especially clearer pathways to legal status for law-abiding immigrants, healthcare reform, curbing waste and fraud, a...
reut.rs
February 10, 2026 at 3:50 AM
one thing that will always be a mystery to me is the number of trump voters who basically thought they were electing george bush again reut.rs/3Mc4h4H
“More achievable.” No, a self growing city on the moon is lot more achievable than a mars trip in 20. Neither is achievable.
Holding my breath for 1 hour is not more achievable than holding my breath for 5 hours.
Holding my breath for 1 hour is not more achievable than holding my breath for 5 hours.
Elon Musk's ambition to one day settle Mars appears to have taken a back seat for a rather nearer and more achievable goal – sending humans to live on the moon. https://cnn.it/46sbAMm
February 10, 2026 at 4:20 AM
“More achievable.” No, a self growing city on the moon is lot more achievable than a mars trip in 20. Neither is achievable.
Holding my breath for 1 hour is not more achievable than holding my breath for 5 hours.
Holding my breath for 1 hour is not more achievable than holding my breath for 5 hours.
Reposted by Bijan Parsia
He murdered 300,000 people last year right here on earth and the toll may climb to 14 million.
Elon Musk's ambition to one day settle Mars appears to have taken a back seat for a rather nearer and more achievable goal – sending humans to live on the moon. https://cnn.it/46sbAMm
February 10, 2026 at 3:12 AM
He murdered 300,000 people last year right here on earth and the toll may climb to 14 million.
Reposted by Bijan Parsia
see how easy it is when they're not as rich and powerful? now keep going
February 10, 2026 at 1:46 AM
see how easy it is when they're not as rich and powerful? now keep going
Reposted by Bijan Parsia
The idea of a bill rendering people deportable for following religious law should make every Jew ill
Tomorrow I will testify before subcommittee of House Judiciary Committee against the proposed "Preserving a Sharia-Free America Act," which would authorize exclusion/deportation of all or most non-citizen Muslim immigrants. My written testimony posted here: reason.com/wp-content/u...
reason.com
February 10, 2026 at 3:47 AM
The idea of a bill rendering people deportable for following religious law should make every Jew ill
Reposted by Bijan Parsia
Thank goodness, the maintenance of a class of people who are raised to think fucking kids, especially the kids of classes who can't fight back and who can be silenced, is an absolute nightmare
Support for British monarchy falls amid Epstein scandal fallout trib.al/D7rkyZL
Support for British monarchy falls amid Epstein scandal fallout
trib.al
February 10, 2026 at 4:10 AM
Thank goodness, the maintenance of a class of people who are raised to think fucking kids, especially the kids of classes who can't fight back and who can be silenced, is an absolute nightmare
Reposted by Bijan Parsia
Quick thread about pitfalls in the Epstein Files, stuff that I'm seeing reported as fact that's somewhere between dubious and outlandish.
February 10, 2026 at 3:56 AM
Quick thread about pitfalls in the Epstein Files, stuff that I'm seeing reported as fact that's somewhere between dubious and outlandish.
Reposted by Bijan Parsia
i think this more than anything is what miller wants. vast detention camps with all of the suffering and death that means.
ICE has now spent over half a BILLION dollars just on purchasing warehouses around the country to convert into detention camps.
If these mega-camps are utilized to the full capacity ICE intends, they'll be the largest prisons in the country, with little real oversight. www.ajc.com/politics/202...
If these mega-camps are utilized to the full capacity ICE intends, they'll be the largest prisons in the country, with little real oversight. www.ajc.com/politics/202...
ICE to begin detaining immigrants inside Social Circle warehouse in April
Homeland Security plans to build warehouse detention facilities in other cities being met with opposition.
www.ajc.com
February 10, 2026 at 3:56 AM
i think this more than anything is what miller wants. vast detention camps with all of the suffering and death that means.
Musk is not a serious to intelligent person. He’s just rich and has a rep. That’s it.
February 10, 2026 at 4:13 AM
Musk is not a serious to intelligent person. He’s just rich and has a rep. That’s it.
I mean you could build a more plausible moon biosphere habitat. Don’t even attempt to deal with construction, vacuum, gravity, etc. We can spot all these. Just build a habitat which contains stuff you could ship to the moon and see how long you can make it last.
February 10, 2026 at 4:13 AM
I mean you could build a more plausible moon biosphere habitat. Don’t even attempt to deal with construction, vacuum, gravity, etc. We can spot all these. Just build a habitat which contains stuff you could ship to the moon and see how long you can make it last.
I remember the Biosphere 2 projects. That actually tried to test closed ecosystems and made some progress but it was tricky.
And this was on earth with like a protective atmosphere. And you could plant stuff.
en.wikipedia.org/wiki/Biosphe...
And this was on earth with like a protective atmosphere. And you could plant stuff.
en.wikipedia.org/wiki/Biosphe...
Biosphere 2 - Wikipedia
en.wikipedia.org
February 10, 2026 at 4:13 AM
I remember the Biosphere 2 projects. That actually tried to test closed ecosystems and made some progress but it was tricky.
And this was on earth with like a protective atmosphere. And you could plant stuff.
en.wikipedia.org/wiki/Biosphe...
And this was on earth with like a protective atmosphere. And you could plant stuff.
en.wikipedia.org/wiki/Biosphe...
I know this is a hoary point but we literally have places on Earth that are enormously less harsh than the Moon which we don’t/cant/won’t settle even to learn about how to do it. And it’s low risk compared to…everything space.
February 10, 2026 at 4:13 AM
I know this is a hoary point but we literally have places on Earth that are enormously less harsh than the Moon which we don’t/cant/won’t settle even to learn about how to do it. And it’s low risk compared to…everything space.
I would be very very surprised if we could build a “self growing city” that was completely underwater on a continental shelf. And that’s literally trivial in comparison. Are there any permanent, inhabited underwater bases anywhere in the world right now?
February 10, 2026 at 4:13 AM
I would be very very surprised if we could build a “self growing city” that was completely underwater on a continental shelf. And that’s literally trivial in comparison. Are there any permanent, inhabited underwater bases anywhere in the world right now?
Just the part on moon dust is totally worth a gander. It’s fascinating stuff and totally worth researching. Like I don’t mind money going into this at all. But we’re not having a “self growing city” in 10 years. No way. No how.
February 10, 2026 at 4:13 AM
Just the part on moon dust is totally worth a gander. It’s fascinating stuff and totally worth researching. Like I don’t mind money going into this at all. But we’re not having a “self growing city” in 10 years. No way. No how.
“Your moon rock building or tunnel might just collapse on you for all we know. Good luck with that.”
February 10, 2026 at 4:13 AM
“Your moon rock building or tunnel might just collapse on you for all we know. Good luck with that.”
“However, destructive testing of real regolith remains limited, relying instead on non-contact measurements and multi-scale simulations, necessitating further research to fully characterize its behavior critical for determining base load-bearing capacity and structural longevity.”
!!!!
!!!!
February 10, 2026 at 4:13 AM
“However, destructive testing of real regolith remains limited, relying instead on non-contact measurements and multi-scale simulations, necessitating further research to fully characterize its behavior critical for determining base load-bearing capacity and structural longevity.”
!!!!
!!!!