GPT?
For me, GPT still means GEC Plessey Telecommunications. who made the System X telephone exchanges, but I guess that they completely disappeared around the turn of the century when they were absorbed into Siemens.
The US Patent and Trademark Office has rejected OpenAI's request to trademark "GPT," which stands for Generative Pre-trained Transformer, the architecture powering its large language models. OpenAI filed to trademark GPT last year, but trademark attorneys didn't approve its application on the grounds that the term was "merely …
> OpenAI, however, reportedly disagreed and said that people wouldn't understand what "Generative Pre-trained Transformer" means.
Exactly, but that's why they didn't need to have used "GPT" as part of their product name in the first place! Virtually no-one outside the field had heard of a "GPT" back then, let alone knew what it stood for, so they could have used an arbitrary and easily-trademarkable marketing name with no risk of confusion.
But that's not the issue.
The issue is that they chose *not* to do that, and instead named their product after a descriptive term already in generic use within the industry. And now they have the problem that the generic term has become strongly associated with their product?
That might be a problem, but it's very much *their* problem and entirely of their own making.
Well, if you have a lot of content you want to search for, an AI chatbot customized on your data could come in handy. Information stored in txt, .pdf, .doc/.docx and .xml formats can be processed.
You mean like Spotlight (OS X, 2005, also running on iPhones under iPhone OS 3, 2009), or Windows Search AKA Instant Search in Vista, 2007, or doubtless countless earlier examples of third party content index or equivalent search systems across those platforms and many others? Or the systems that even start to use dramatically more efficient, domain-optimised on-device ML models for things like object detection in photographs and videos?
I mean sure, they don't need an RTX 30xx and gigabytes of RAM - it'll even run on an old phone, for example, and they handle a dramatically wider and extensible set of file formats - but the results are guaranteed accurate, instead of risking hallucinations. And where's the fun in that?!
/s
Hell, in the mid-1990s Bradley Rhodes introduced (and open-sourced) his Remembrance Agent, which combined the Savant word-stemming-and-indexing system1 with an Emacs-based UI. It continually searched based on whatever you were typing in Emacs and displayed the list of the top N relevant documents down at the bottom of your Emacs window; you could switch to the list and open any of those with a couple of keystrokes.
I'd call that a much, much, much better search UX than "conversational" will ever be. (Agrawala lays out some of the reasons in his ACM talk "Unpredictable Black Boxes are Terrible Interfaces". More generally, though, when searching for information, specificity matters, which means precision and accuracy matter, which is why we have librarians.)
And most use cases for LLMs position them as competitive cognitive tools, which means they're encouraging intellectual laziness, among other problems.
Searching is a terrible use case for LLMs.
1Using the SMART algorithm, which was published in 1971.
Gemini will record and store every conversation for three years? Seems like "Miranda warnings" should be required at the start of all AI chat sessions... "Anything you say can and will be used against you."
I know the "Miranda rights" are a US thing, but I suspect enough US TV has leaked across to the UK that the reference will be understood. Apologies if not.
I avoid the chat "help" functions on websites like I avoid the plague. Why people think it is better to communicate using a typed-out back and forth exchange than to communicate by an interactive voice conversation is beyond me. I get that texting is a cool asynchronous partial-attention communications tool, and I use it that way myself. But when I need to get a problem solved, and I'm reaching out to get help, I don't want asynchronous and I don't want partial-attention. I want full-on interactive attention with voice inflection and all the other queues provided by a live audio stream, along with the feeling of a personal connection. Even if the chat session endpoint claims to be human, one can never be too sure.
Microsoft have also admitted to storing and analyzing LLM chat sessions, as Schneier noted yesterday.