Hacker Newsnew | past | comments | ask | show | jobs | submit | hatenberg's commentslogin

It’s explicitly part of Chinese science and technology strategy to think outside the box and it’s what’s pushing them forward on areas like semiconductors as well.

Given the massive advantage in talent they’ve built up while the Us reverts to Drill Baby Drill we know how this ends.

Eventually the Us with push for atmospheric dimming to “fix” the negative externalities of their approach which had the nice side effect of degrading solar ….

FMl


You’re being downvoted but i don’t think you’re entirely wrong. China has been pursuing some stuff that the western world had essentially abandoned, getting interesting wins (eg: thorium reactors).


He's wrong in the "why" tho. It's not that they must think outside the box, it's that they _must_ not all focus on a single point of research. I'm pretty sure that they are also pursuing popular research topics, because it would be pretty bad if they fall behind for not doing the obvious.


Solar and batteries will keep getting cheaper, its going to get harder to maintain fossil fuel use under those market conditions.


Competition is great.


The police is to blame for trying to enforce the law, it makes criminals innovate is exactly the kind of take I come here for.


I mean that's exactly what is happening right now. Companies are buying the assets of companies instead of the companies themselves so they don't have to worry about being sued which is a timely and costly process that has become very unreliable due to how much politics are involved in the process now.


Big Tech: We shouldn’t pay, each individual piece of content is worth basically nothing.

Also Big Tech: We added 300.000.000 users worth of GTM because we trained in the 10 specific anime movies of Studio Ghibli and are selling their style.


The funny thing is that style is not copyrightable.


Exept it's a rectangle with 4 rounded corner.


Patents are not the same thing as copyright.


"Pretraining data us worth basically nothing."

(Raises $10 billion based on estimated worth of the resulting models.)

"We can't share the GPT4 prettaining data or weights because they're trade secrets that generate over a billion in revenue for us."

I'll believe they're worth nothing when (a) nobody is buying AI models or (b) AI companies stop using the copyrighted works to train models they sell. So far, it looks like they're lying about the worth of the training data.


More layers of abstraction, more!


Nah it’s just that Zuck watched the Barbie movie and realized Soace Karen was getting entirely too much limelight and declared a Year of Masculinity


Calling someone Karen is a misogynist slur, and calling a man by a woman's name without consent is doubly a misogynist slur.


Ok Karen


Even with RAG you’re bounded at some 93%, it’s not a panacea.


How are you bounded? When you can easily check the sources? Also you act as if humans without LLMs have a higher success rate?

There is an entire “reproducibility crisis” with research.


False equivalence. Humans fail in predictable ways that can be mitigated though process and training.

Try training an LLM.


How are the results any different? You can’t completely trust human output. There are all sorts of cognitive biases that come into play.


Why purchase he’s doing AI training /s


This is a classic.

We put a human mask on a machine and are now explaining it’s apparent refusal to act like a human with human traits like deception.

It’s so insulting you have to wonder if they are using an LLM to come up with such obvious narrative shaping.


Gotta love a new age of slavery where human drones in the colonies are serving us without us having to look into their faces.

Now the anti immigrant push makes perfect sense, it’s priming demand for the slave labor matrix


The plan was regulatory capture and defanging. It worked.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: