I'm a novice in this area, but my understanding is that LLM parameters ("neurons", roughly?), when processed, encode a probability for token selection/generation that is much more complex and many:one than "parameter A is used in layer B, therefore suggest token C", and not a specific "if activated then do X" outcome. Given that, how would this work?
The key part of the article is that token structure interpretation is a training time concern, not just an input/output processing concern (which still leads to plenty of inconsistency and fragmentation on its own!). That means both that training stakeholders at model development shops need to be pretty incorporated into the tool/syntax development process, which leads to friction and slowdowns. It also means that any current improvements/standardizations in the way we do structured LLM I/O will necessarily be adopted on the training side after a months/years lag, given the time it takes to do new-model dev and training.
That makes for a pretty thorny mess ... and that's before we get into disincentives for standardization (standardization risks big AI labs' moat/lockin).
I'm curious why this was downvoted--I'm not complaining or trying to go against HN guidelines; I'm genuinely unclear as to why the first-party source for the article clarifying the question in GP was marked dead. Bad actors? Misinterpretation? Other?
No idea, I thought it was a valid question and we go to great lengths in our methodology for this reason. The audits we supply for enterprise are highly specific as to cookie purpose for this reason: https://webxray.ai
We've also made progress as a species towards banning and reducing other things that in-group upsides and really bad externalities: off-the-shelf sale of broad system antibiotics; chattel slavery; human organ trafficking; some damaging recreational drugs.
The prohibitions aren't perfect, of course (and not without their own negative externalities in some cases). But all of those things are much more accessible to people than nuclear weapons, and we've still had successes in banning/reducing them. So maybe there's hope yet.
This is an entertaining (and often exasperating) decades-old trend in competitive U.S. college debate, as well.
A common advantageous strategy is to take the randomly-selected topic, however unrelated, and invent a chain of logic that claims that taking a given side/action leads to an infinitesimal risk of nuclear extinction/massive harms. This results in people arguing that e.g. "building more mass transit networks" is a bad idea because it leads to a tiny increase in the risk of extinction--via chains as silly as "mass transit expansion needs energy, increased energy production leads to more EM radiation, evil aliens--if they exist--are very marginally more likely to notice us due to increased radiation and wipe out the human race". That's not a made-up example.
The strategy is just like the LessWrongers' one: if you can put your opponent in the position of trying to reduce P(doom), you can argue that unless it's reduced to actual zero, the magnitude of the potential negative consequence is so severe as to overwhelm any consideration of its probability.
In competitive debate, this is a strong strategy. Not a cheat-code--there are plenty of ways around it--but common and enduring for many years.
As an aside: "debate", as practiced competitively, often bears little relation to "debate" as understood by the general public. There are two main families of competitive debate: one is more outward-facing and oriented towards rhetorical/communication/persuasion practice; the other is more ingrown and oriented towards persuading other debaters, in debate-community-specific terms, of which side should win. There's overlap, but the two tend to be practiced/judged by separate groups, according to different rubrics, and/or in different spaces or events. That second family is what I'm referring to above.
I can't point out where I draw the line clearly but here's one different I notice:
A recommendation can be both a thing and an action. A piece of text is a recommendation and it does not matter how it was created.
Assistance implies some parity in capabilities and cooperative work. Also it can pretty much only be an action, you cannot say "here is some assistance" and point to a thing.
Even adjusted for maternity and career entry/exit differences, the gender pay gap is still big and real. And while there is overlap and outliers, many more women-dominated industries are at lower pay segments (and with fewer benefits) than industries dominated by men.
Also worth noting that several studies have shown pay differentials to be highly correlated with women being less likely to negotiate compensation or ask for less.
Do you think your comment was any of those things?
Why are you free from the bounds of courtesy, but fall back on them the moment someone responds in kind?
Do you really believe that women cannot arc weld?
Do you really believe that women cannot scuba dive?
Do you really believe that women cannot lift 25kg or take out the trash?
If you do, then it's rather unkind and incurious to cast half of the human population as so fundamentally incapable, and you should bring evidence to support your assertion.
If you don't actually believe it, then the substance of your comment is unkind snark, no?
Your original comment is just a warmed-over version of "women wouldn't last a day in a real man job"
We've been hearing that repeated ad infinitum as women have continued to prove the assertion wrong.
For the record though, I don't think anyone should do a job that has a 1/15 fatality rate. Men shouldn't be cannon fodder either.
Next you'll be telling me that male US farmers roll up several thousand yards of fencing by hand to move it rather than have female jillaroos spool it up on a hydraulic spindle.
Underwater welding does pay extremely well. But to break into that field needs a commercial diving license and lots of time/skill at saturation diving. Basically, it is easier to train a diver to weld than it is to teach a welder to be a commercial diver. Many (US) entrants are ex-SEAL. There's about 20-25k such people worldwide and the death rate is about 1 in 15. Career ending injuries are about the same. If you need nightmare fuel: "Byford Dolphin".
The article is pretty clear, Women are getting most of the new jobs because they're in fields that Men largely don't try to enter (ex. Teaching).
Like whats the big initiative to increase the amount of Women with a Masters of Education? I've heard of a bunch for STEM but Men still dominate that field but that field is growing slower than other Women dominated ones so it's a non-sequitor.
We do?
reply