Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Don Y <blockedofcourse@foo.invalid>
Newsgroups: sci.electronics.design
Subject: Re: The end of stackoverflow?
Date: Fri, 10 May 2024 01:29:18 -0700
Organization: A noiseless patient Spider
Lines: 17
Message-ID: <v1klt7$18ifi$2@dont-email.me>
References: <v1kgcf$3dc4$1@solani.org> <la5vu5FojftU1@mid.individual.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Fri, 10 May 2024 10:29:28 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="34557ccb4cf0a0dfc074d582ac51803c";
	logging-data="1329650"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19fuObjR6G9YbztrxAeSPsq"
User-Agent: Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.2.2
Cancel-Lock: sha1:PgGX3+4KAoG1xy/VbnalEdlCw2U=
Content-Language: en-US
In-Reply-To: <la5vu5FojftU1@mid.individual.net>
Bytes: 1851

On 5/10/2024 12:16 AM, Sylvia Else wrote:
> One often has to trawl through a number of suggested solutions, either because 
> most of them are wrong (or at least wildly apocryphal), irrelevant, or because 
> the same or similar symptoms can have many different underlying causes.
> 
> I have to wonder whether a language model is really up to the task of filtering 
> out the dross, while keeping the important parts.

Patterns (repeated) in answers are reinforced.  So, outliers tend to
not influence the model, as much.

E.g., Carlin (?) did a bit in which he uttered something like, "Here's a
sentence no one has ever said before..."  You, thus, wouldn't expect an
AI to come up with such a sentence in "normal use" because its weights
are so low.