Lerc

53 ms·

901.
by Lerc 1y ago
It's not often you read a title like that and expect it to pan out, but from a quick browse, it looks pretty good. Now I just need a time-turner.
902.
by Lerc 1y ago
>The issue is NOT lack of such database per se, the issue is lack of impartial oracle used to build such database. You don't need a perfectly impartial oracle, sure that would make the task much simpler. You build examples from i
903.
by Lerc 1y ago
Doesn't really matter. All of the gains made before any funding collapse will exist. If you look at the flow of papers coming out right now, there are a massive number of intriguing ideas that will not get a chance to be included in t
904.
by Lerc 1y ago
I kind of hope that they will get there. I don't know that they will, but I'm hopeful. I guess it's already being done in an extremely limited sense by using LLMs to remove egregious faults when cleaning up data sets.
905.
by Lerc 1y ago
I have been looking at things like this and wondered if there is scope for an open source project to design boards that provide power and data access to a variety of common old laptop panels (and keyboard+touchpads) along with a holder for
906.
by Lerc 1y ago
Well I cannot address the accuracy of your statement on UPFs unless I assume the term UPF is synonymous with the Nova Classification, which is much of the problem. If it is synonymous then the term is bound to a classification system and th
907.
by Lerc 1y ago
That's why I stated this case as an opportunity. To be able to do this you would need to have a set of examples where sometimes the dominant narrative is incorrect. This article represents one of those cases. Identifying more would
908.
by Lerc 1y ago
On that I disagree. LLMs are not simple Markov chains. They may fail at a lot of logical tasks, but I don't think that is the same as exhibiting no logic. Getting even slightly respectable performance on the ARC-AGI test set, I think
909.
by Lerc 1y ago
I think the issue with them making assumptions and failing to properly diagnose issues comes more from fine-tuning than any particular limitation in LLMs themselves. When fine tuned on a set of problem->solution data it kind of carries
910.
by Lerc 1y ago
The LLM aspect of this, I think shows both a common weakness and an opportunity. If you suspect something is a commonly held misconception, frequently asking a LLM about it is close to useless, because the abundance of text repeating the mi
911.
by Lerc 1y ago
Surely bad actors leave a fairly clear data trail. Are there no analytics being used to track this sort of behaviour? Much of the scale of this comes from being able to do it with impunity. If bad actors were exposed, even after the fac
912.
by Lerc 1y ago
There's an issue here with classification and harm. You can only say that UFPs are harmful if the definition used when measuring harm matches the definition used elsewhere. It seems all to easy to attach the harm to the label then s
913.
by Lerc 1y ago
So is the same item counted as UPF when the USDA supplies it but not otherwise, or does it mean the USDA supplying it grant the category of item non-UFP status. Does this count if they supply it from a vending machine on their premises :-)
914.
by Lerc 1y ago
so is that “ultraprocessed food” or “UPF” means (any food or beverage that contains a substance described in paragraph (2) and either high amounts of saturated fat, sodium, or added sugar, as described in subparagraph (A) of paragraph (3),
915.
by Lerc 1y ago
Those seem like particularly bad reasons. I'm not sure if they are the arguments that the author has been given or if that's what he perceived those arguments to be. My take on it is to remember that the people you are talking t
916.
by Lerc 1y ago
Is there a need for a wiki to collect instances of these stories? Having a place to check what claims are actually being made rather than details drift like an urban legend. I tried to determine the origin of a story about a family being
917.
by Lerc 1y ago
With Gemini I have found some weird issues with code gen that are presumably temperature related. Sometimes it will emit large block of code with a single underscore where it should be a dash or some similar very close match that would mak
918.
by Lerc 1y ago
It also implies that either the blood was taken within 30 minutes of inhalation or they are measuring something else entirely. Post mortem diffusion can happen, in either direction. Unless the individual died while smoking, there would have
919.
by Lerc 1y ago
To answer the question literally, I don't think it is possible to know. There are a lot of elements at play. - There are valid concerns about the dangers of AI. - There are hysterically overblown claims about the dangers of AI - There
920.
by Lerc 1y ago
I have always wanted to give this a try. I have this daft idea of implenting a system logically as an n-dimensional cube and physically on two facing surfaces (even and odd parity processor addresses) with laser connections between the pr
921.
by Lerc 1y ago
' If you flipped a coin 100 times and all you got are heads you ' By starting the sentence with if, you are selecting the occurrence to look at. If you said I am about to look at the results of this coin toss that happened yesterd
922.
by Lerc 1y ago
I think there is a bit of disconnect between people knowing what is possible and what people fear might be doable. It's entirely possible that there are good non technical reasons for believing who was behind this while being technical
923.
by Lerc 1y ago
I felt like it was more likely to be a complete absence of a human in the loop.
924.
by Lerc 1y ago
Nitrogen oxygen at one atmosphere is a lifting gas on Venus. You could live in an enormous zorb if you can keep it sealed.
925.
by Lerc 1y ago
Not the argument. The elimination of life on earth can happen due to non human causes. An impact of the sort that created the moon would do the job.
926.
by Lerc 1y ago
Are there? Most places on earth have an established environment. There are things living in some very hostile to human areas.
927.
by Lerc 1y ago
What the developer learns from the experience of AI programming is more down to the attitude of the individual than that. Here's an example from my perspective. Recently while developing a way to output a PAL video signal from two digi
928.
by Lerc 1y ago
I would be interested to see how capable it would be at going against cultural standards to preserve accuracy. It's very hard for humans to imagine some mundane things when they go against their expectations. I would imagine AI would s
929.
by Lerc 1y ago
With floats getting smaller and smaller in ML, It's hard to imagine anyone failing to learn this as one of their early experiences in the field. The focus should not be on the possibility of error, but managing the error to be within a
930.
by Lerc 1y ago
My intuition is that this would include every 4 state (and fewer) machine that has a fifth state that is never used. For every different configuration of that fifth state I would consider the machine to be essentially the same.

More ›