Calvin

AI enthusiast, commercial fisherman, third and final distinguishing feature.

[Calvin @CalvindeFox]

@maiab Because people won't pay. Only something like 2% of US households are paying for AI right now I hear. Blows my mind

[Calvin @CalvindeFox]

@bryancsk You can't adjust the language with an LLM?

[Calvin @CalvindeFox]

@micsolana It is slavery

[Calvin @CalvindeFox]

@MatthewBerman Agreed!

[Calvin @CalvindeFox]

@AndrewCurran_ Hmm 🤔 I hope this isn't like the Fable bio restrictions and now anything "mean" gets flagged. I never abuse Claude, but I say terrible things about other people in his presence and ask him what he thinks.

[Calvin @CalvindeFox]

@rad_niazadeh Why not?

[Calvin @CalvindeFox]

Hey thanks for laying it out for me. Im enjoying the exchange.

Im hearing the risks as:

1. Potential loss of skill among experts if machines do the cognitive labor
2. Almost certain devaluation of papers published which is the backbone of academic status and funding system. Potentially ruining the community along with the system
3. Scooping mathematicians is like "salting the earth" in their research area and demoralizing to the people working that terrain.
4. Seems like cheating

My thoughts

1. Skills will revalue. Some will fall, some will rise and many new ones will appear. LRMs when used "skillfully" offer almost unlimited skill building potential for the motivated individual at a price middle income maybe even low income countries soon can afford. Everyone gets a facsimile of a PhD advisor not only the privileged.
2. True. Academia will need to figure out a new system of merit. From what I hear many academics weren't too thrilled about the publish or perish regime to begin with. Smart people, real problem.
3. I have a hard time understanding this view. If I was working on a problem I would start with a lit review and I would hope that would include ALL discovered information. Then I'd look for the holes and focus my attention there. Ideally I'd have access to the SOTA model as an assistant, but second best would be its results published openly. Tons of new small discoveries and maybe some large ones are rocking the math world. It's uncomfotable.
4. You can't cheat in discovery. Either you've discovered something or you havent. Cheating the publication system is real, but let's fix the system not encourage mathematicians not to work with the single most powerful math tool/assistant ever invented.

The way I see it Mathematicians should be advocating for MORE tokens spent on THEIR problems as hard as they can because there's only so much juice in the system. The transition is rough. Software engineers are divided too. Some are complaining because they feel disempowered managing agents instead of practicing their art, some aren't sleeping because all of a sudden they have superpowers and they can't stop building. I think the latter is the healthier approach assuming one eventually finds the time to sleep.

[Calvin @CalvindeFox]

@yukikaze85 @Tyler_A_Harper Many set netters teach. The seasons line up perfectly

[Calvin @CalvindeFox]

@passtheadjoint @elsleightholm 😂 touché, I did mean numerate. That's priceless.

Maybe you could explain the security concerns or at least one of them. So far all you have said is that I dont understand. Dont understand WHAT? What is the risk of attempting to solve open math problems with LRMs?

[Calvin @CalvindeFox]

@chicfryrice It worked!

[Calvin @CalvindeFox]

@nicolas_oyarceg @elsleightholm I can assure you that you never know

[Calvin @CalvindeFox]

Spanish has relatively transparent orthography with only a handful of exceptions. Finnish, Turkish, Serbian, Bosnian, Romanian, Hungarian, Czech, Indonesian and Swahili are all very transparent. In fact, English scores among the least transparent of the major European languages so difficult and opaque that we made a sport out of guessing how a word is spelled. It's ridiculous. Almost as ridiculous as our system of weights and measures. We do this to ourselves and most people don't even know that there is a better way.

[Calvin @CalvindeFox]

@passtheadjoint @elsleightholm Or worse, that the prompts should never have been run

[Calvin @CalvindeFox]

@passtheadjoint @elsleightholm As long as nobody is trying to keep results secret from the world I have no objection. I understood the situation to be the committee was recommending the findings not be released

[Calvin @CalvindeFox]

@bendreyfuss Interesting idea, but how do we know they would care? They may be suicidal

[Calvin @CalvindeFox]

@cremieuxrecueil We are the hyper social species. We move together

[Calvin @CalvindeFox]

@TheAlexSylvian In one dollar bills? At that rate it's not the money Im concerned about it's the volume. Finding the original dollar before suffocating probably needs to happen before day 30 if I remember the parable right

[Calvin @CalvindeFox]

@Sonderisfon Take a piss

It's how every day is started