181–190 of 297 posts
Earlier quoted context omitted.
why don't you just use chatGPT? from what i know it's running GPT3.5 and it's not that different (at least in terms of code quality)
When my 25 queries per 3 hours runs out I don't use openai at all. That's how bad chat gpt is in comparison to gpt4 in my use cases.
I don't understand how this can be considered a technical report. No information on model architecture, distributed training methodology, or optimizations. The "Training dataset" section is a pathetic 0.5 pages long. Come on, Google.
LLM is going to make money, a lot of money, nobody is going to give away their secret sauce for free.
Prepare for the landscape to get really ugly and really soon. Maybe we will witness some epic legal battle around big techs.
Earlier quoted context omitted.
this is about the time that i expect sites to begin returning intentionally corrupt/incorrect/perhaps outright garbage (subtle or not, probably better subtle so they don't realize it until it's far too late) data in order to intentionally poison enemy wellscraping. where "ethics" dissolve into the inherent raw cannibalistic laws of capitalist ventures. then you can sell them back the TBs they scraped at a 1000x marku…
You might be right. What a dystopian future that will be. Make a few requests too many and the webserver might think you're scraping data so it gaslights you into reading bullshit.
Earlier quoted context omitted.
Btw I've arrived at a different interpretation of the "Open" in OpenAI. It's open in the sense that the generic LLM is exposed via an API, allowing companies to build anything they want on top. Companies like Google have been working on language models (and AI more broadly) for years but have hid the generic intelligence of their models, exposing it only via improvements to their products. OpenAI bucked this trend an…
> Btw I've arrived at a different interpretation of the "Open" in OpenAI. I don't understand why people have to keep trying to wrap their head around the word 'Open' in OpenAI. If you ever saw a commercial like a product has a 'great new taste' but then you tried it and it tasted bad, would you twist yourself into knots trying to understand how you went wrong in your interpretation of 'great'? No that's ridiculous. S…
Earlier quoted context omitted.
Btw I've arrived at a different interpretation of the "Open" in OpenAI. It's open in the sense that the generic LLM is exposed via an API, allowing companies to build anything they want on top. Companies like Google have been working on language models (and AI more broadly) for years but have hid the generic intelligence of their models, exposing it only via improvements to their products. OpenAI bucked this trend an…
Guess they should rebrand as AvailableAI then...
Earlier quoted context omitted.
Most of the GPT-4 benchmarks from their report were things like AP tests or leer code scores. Which aren’t benchmarks that can be compared by a different set of researchers as you don’t know the constituent parts of the test to run
GPT-4 report has MMLU score, which is believed to one of the most important metric for question answering task. GPT-4 MMLU score is slightly higher than PaLM 2(86 vs 81). Google didn't compare it in with PaLM 2 in this paper.
Earlier quoted context omitted.
I've asked "are you using palm 3": It said: I am using the Palm 3 model. Palm 3 is a large language model... Don't believe it :) Also, In the technical report, It mentions multiple languages, I've asked in Turkish which was supposed to be supported, but wasn't able to answer. Even if its PaLM 2, its hard to trust to the model itself.
I asked it "are you using the palm 420 language model or the palm 2 language model?" It said "I am not using either the Palm 420 language model or the Palm 2 language model. I am using a different language model called Bard, which is a large language model from Google AI." Perhaps the people at Google saw this and made a manual correction? Hard to say, black boxes and all...
Bard is very bad about this, it lies with the slightest encouragement. Try "compare and contrast captain planet and paptain clanet".
Earlier quoted context omitted.
Is it better than bing or phind though? Why would I use it over bing?
It isn't Edge-specific which is good and I find it faster than Bing. Phind is way better than Bard, but verbose. I still find ChatGPT my first port of call. GPT-3.5 is blazing fast and very useful.
"The PaLM 2 pre-training corpus is composed of a diverse set of sources: web documents, books, code, mathematics, and conversational data" I really want to know more about the training data. Which web documents, which books, code from where, conversational data from where?
I fully expect Discord to be a data source, if not already, then for a future version. I also expect that the only way the general public would ever find this out is via whistle-blower.