upvote

points

by dakolli3 hours ago |
comments
upvoteby handoflixue2 hours ago|
next
[-]
> LLMs are literally stochastic by nature and can't be relied on for anything critical

Ahh, yes, unlike humans, who are completely deterministic, and thus can be trusted.

reply
upvoteby steveBK1231 hours ago|
parent|
next
[-]
Humans can be governed by rules with consequences and replaced with individuals with a appropriate level of risk taking / rule following for the role.
reply
upvoteby somewhatgoated1 hours ago|
parent|
next
[-]
Rules and consequences seem to apply to humans in a similar way as prompts and harnesses govern LLMs. The greater the level of power a human possesses the less they are governed by these restraints, this doesnt apply to LLMs so at least in that aspect they are an improvement. But yea we can’t really punish or inflict pain on them - this seems like a problem
reply
upvoteby steveBK1231 hours ago|
parent|
next
[-]
I think a simpler model is variety.

There are billions of people, you can interview/hire/fire until you get the right match.

There are 2? frontier LLM providers. 5? if you are more generous / ok with more trailing edge.

Everyone thought OpenAI was great, until Claude got better in Q1 and they switched to Anthropic, and then Codex got better and a good chunk moved back to OpenAI.. Seems kind of binary currently.

reply
upvoteby handoflixue1 hours ago|
parent|
prev|
[-]
Why does it matter if you can inflict pain on them? Is that normal and acceptable in your line of work?
reply
upvoteby wild_egg1 hours ago|
parent|
[-]
Being able to fire someone, thus causing potentially significant hardship, is considered quite normal and acceptable in most lines of work.
reply
upvoteby somewhatgoated42 minutes ago|
parent|
[-]
Yea I didn’t mean actual physical violence but rules need painful consequences in some way to be meaningful?
reply
upvoteby GigaDingus1 hours ago|
parent|
prev|
next
[-]
Which has, famously, been a great consolation for people who suffered the consequences of human failure in the past
reply
upvoteby handoflixue1 hours ago|
parent|
prev|
next
[-]
That seems like it applies just fine to LLMs as well: You can replace an LLM with a different model, different prompts, etc. for the appropriate level of risk taking. Rule following is even easier, given you can sandbox them.
reply
upvoteby steveBK1231 hours ago|
parent|
[-]
Theres at best a handful of frontier models vs billions of people and millions of SWEs.
reply
upvoteby someguyiguess1 hours ago|
parent|
prev|
[-]
You clearly have never met a human
reply
upvoteby steveBK1231 hours ago|
parent|
[-]
If you cannot get humans to do roughly what you want as a manager, good luck with LLMs.
reply
upvoteby hansmayer1 hours ago|
parent|
prev|
[-]
Wow, such a nasty view to hold. What's next, the Altman's bullshit argument about "all the food" that the humans need to grow up and develop brain ? Humans are intelligent. Humans can generalise and invent new concepts, ideas and art. LLMs are none of that.
reply
upvoteby keybored2 hours ago|
prev|
[-]
What is the ad hoc fallacy? From googling I didn’t find any convincing definitions (definitions that demonstrate that it is a logical fallacy).
reply
upvoteby jibal2 hours ago|
parent|
[-]
https://finmasters.com/ad-hoc-fallacy/

> Ad hoc fallacy is a fallacious rhetorical strategy in which a person presents a new explanation – that is unjustified or simply unreasonable – of why their original belief or hypothesis is correct after evidence that contradicts the previous explanation has emerged.

https://cerebralfaith.net/logical-fallacy-series-part-13-ad-...

> An argument is ad hoc if its only given in an attempt to avoid the proponent’s belief from being falsified. A person who is caught in a lie and then has to make up new lies in order to preserve the original lie is acting in an ad hoc manner.

It should be clear why the ad hoc fallacy is a fallacy.

reply