Simo Hosio

Optimist Professor in Computer Science & Engineering / 🇫🇮 & 🇯🇵 / Obliterate your academic obstacles 😤

[Simo Hosio @simohosio]

@CoachDanGo it's fun to party for 5 years tho -- far from useless! well, the __degree__ is different than the social experience, you're right in that...

[Simo Hosio @simohosio]

@xriskology hahah i feel so alienated from most discussions on how "it's all a scam" as in finland it was/is idd free (well, for eu citizens..)

[Simo Hosio @simohosio]

You know what? AI writing IS NOT poisoning academia.

The sick incentives poisoned it decades ago. AI is just making everyone do what they are told they should do to succeed: make the number go up.

And it also worked for 2-3 years. Now I already see the time window closing. All venues are DROWNING in submissions. Of course, the crappy venues don't have to even care, they'll just publish everything.

I've also been leaning heavily on Claude Code lately for most of my work. Great tool. Teaches me a lot. But also makes a ton of judgment mistakes, which I'm most likely capable of catching, but someone who's just getting started most likely not.

It's a miracle maker. But it won't save you.

In the end, it all boils down to just managing yourself. Putting enough time on the right things. Learning the new tools, reading the research, developing your thinking, taking the time to write down your thoughts (not just work, I literally mean being reflective of your life and work and days), and trying to focus on good kind of problems. Not maybe just the kind that "get published" but that might even have a good impact on the world.

( If you want a quickstart to my philosophy on that front... well you know where to find it: https://t.co/GagGgjufQK )

[Simo Hosio @simohosio]

This ^ is a GREAT pointer to a less-discussed direction imo. Everyone is pretending the function of academia (higher edu), work, etc is just "get more done" or "learn as efficiently as possible"

There's a human element in all this. A purpose for life, as a kind of byproduct of doing all that. And it's largely ignored in the discourse.

[Simo Hosio @simohosio]

@AndieWinnipeg If my university doesn't explicitly ban it, why would I care what it says?

All I care about is the result.

Not saying AI is always making it better, but it's really the only thing I optimise for with grant applications.

[Simo Hosio @simohosio]

@mattbencole genuinely curious, what makes us paying for a service we use daily "sad"

[Simo Hosio @simohosio]

I create much better research and am much more productive with LLMs, but lose the sense of discovery and some deeper ownership.

It's like working with a hyper-productive, very broadly knowledgeable Dr. know-it-all.

Then again, the aspect I most love about academia --- just creating things joyfully in a sandbox (a metaphor I once got scolded live, real-time, in a tenure position interview of all places lol) --- is now on steroids.

Just printing papers, easy.
Doing good work, still hard.
Funding, harder than ever.

Hard doesn't mean bad.
Hard just means it's worth a good shot.

[Simo Hosio @simohosio]

@MagnusPharao haha yes i know the feeling :D

[Simo Hosio @simohosio]

@Marc__Watkins @Tyler_A_Harper yeah but microsoft's tools are so bad nobody will actually use them - that's why they can give them for free

[Simo Hosio @simohosio]

@decadentquill I'm with you, yet I fear that whoever PAYS currently for that...grueling work... might not be, when they find out "number go up" with AI-vomit too...

[Simo Hosio @simohosio]

@bodenlosig Uh oh yeah this sucks. Sometimes I wonder how do people send that crap, but could it really be they can't tell?

[Simo Hosio @simohosio]

@brecht_dp sounds more like a skill issue (bad ai use)

[Simo Hosio @simohosio]

@RexDouglass :D deep cut

[Simo Hosio @simohosio]

@jennfrey The fraud is already here. The post-hoc integrity checks is what will really bite thousands of academics in the butt, years from now (when they are faculty) -- it's going to be ugly

[Simo Hosio @simohosio]

@EraOccidental @jennfrey they dont have time to even review these days, and you'd add interviews?

[Simo Hosio @simohosio]

@SaintOurania @jennfrey Most don't publish the prompts tho

[Simo Hosio @simohosio]

@jennfrey Nah

[Simo Hosio @simohosio]

@lianda_edu hahah i can assure they always don't

( but yeah the idea of prompting a paper out of existing references just because...is repulsive )

[Simo Hosio @simohosio]

@paulnovosad it's also not like you can simply forecast "this guy will have impact" and the other one not. you hire people who are good at doing their work. if it happens to lead to actual impact vs. a pile of "meh" papers, great! if not, well, we tried.