<div dir="ltr"><div>I was at a talk recently by the owner of AVG, Norton etc (can't remember the company name).  He said that there is a lot of evidence now that malicious actors are poisoning the LLM's, creating thousands of websites with incorrect information (disinformation), bad use of language and are also targetting the training for cyber security AI.  Use such open tools with caution... </div><div><br></div><div><div dir="ltr" class="gmail_signature" data-smartmail="gmail_signature"><span></span><span></span>Nick Tudor<div>Tudor Associates Ltd</div><div>Mobile: +44(0)7412 074654</div><div><a href="http://www.tudorassoc.com" target="_blank">www.tudorassoc.com</a></div><div><img src="http://www.tudorassoc.com/wpimages/wpb4e71a5c_0f.jpg" width="200" height="40"></div><div><font color="#00144d" face="Arial, Helvetica, sans-serif" size="1"><b><br></b></font></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><b><span></span><span></span>77 Barnards Green Road</b></font></span></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><b>Malvern</b></font></span></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><b>Worcestershire</b></font></span></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><b>WR14 3LR</b><strong><br>Company No. 07642673</strong></font></span></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><strong>VAT No:116495996</strong></font></span></div><div><span style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif"><font size="1"><strong><br></strong></font></span></div><div><strong style="color:rgb(0,20,77);font-family:Arial,Helvetica,sans-serif;font-size:x-small"><a href="http://www.aeronautique-associates.com" target="_blank">www.aeronautique-associates.com</a></strong>
</div></div></div><br></div><br><div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Wed, 16 Jul 2025 at 17:05, Derek M Jones <<a href="mailto:derek@knosof.co.uk">derek@knosof.co.uk</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Peter,<br>
<br>
>> token prediction.<br>
> <br>
> A request for clarification. Are you referring to October 2022 and the release of ChatGPT?<br>
> <br>
> As I mentioned, LLMs "arrived" between one and two decades ago in computational linguistics. We were using them in the <br>
<br>
People have been doing stuff with word sequence probabilities since<br>
Shannon's famous 1948 paper "A Mathematical Theory of Communication"<br>
<a href="https://web.archive.org/web/20090216231139/http://plan9.bell-labs.com//cm//ms//what//shannonday//shannon1948.pdf" rel="noreferrer" target="_blank">https://web.archive.org/web/20090216231139/http://plan9.bell-labs.com//cm//ms//what//shannonday//shannon1948.pdf</a><br>
<br>
When did LLMs arrive on the scene and when does a tool<br>
become an LLM?  I will leave others to argue over this.<br>
<br>
The 2021 paper you cited<br>
"On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?"<br>
uses the term language model (LM) and does not use the term LLM (but<br>
does talk about "... ever larger language models,").<br>
Also the term does not appear in the famous 2017 paper<br>
"Attention Is All You Need"<br>
<a href="https://arxiv.org/abs/1706.03762" rel="noreferrer" target="_blank">https://arxiv.org/abs/1706.03762</a><br>
<br>
-- <br>
Derek M. Jones           Evidence-based software engineering<br>
blog:<a href="https://shape-of-code.com" rel="noreferrer" target="_blank">https://shape-of-code.com</a><br>
<br>
_______________________________________________<br>
The System Safety Mailing List<br>
<a href="mailto:systemsafety@TechFak.Uni-Bielefeld.DE" target="_blank">systemsafety@TechFak.Uni-Bielefeld.DE</a><br>
Manage your subscription: <a href="https://lists.techfak.uni-bielefeld.de/mailman/listinfo/systemsafety" rel="noreferrer" target="_blank">https://lists.techfak.uni-bielefeld.de/mailman/listinfo/systemsafety</a><br>
</blockquote></div>