alpark3

8 ms·

31.
by alpark3 3y ago
You have to do some intense prompt-engineering with the system prompt. The model considers the system to be trusted (roughly), some of the early 2023 exploits with ChatGPT still work if you do it on the system prompt.
32.
by alpark3 3y ago
XLF is a financial sector ETF. JPM is a huge bank. If JPM does badly, XLF should probably do badly, and vice versa. Then it stands to reason that if JPM falls below XLF, it will soon go back up to where XLF is. This is relative, so the same
33.
by alpark3 3y ago
It's worth noting NLN Holdings is a joint venture of Jump and Virtu, also some big names within the HFT space.
34.
by alpark3 3y ago
Coding? Your art is one of the most directly financially valuable skills on the planet and you're comparing yourself to starving musicians?
35.
by alpark3 3y ago
Here's my experience: me > what is the millionth prime? Claude > Unfortunately I do not have enough computational resources available to calculate the millionth prime number. Finding prime numbers gets exponentially more difficul
36.
by alpark3 3y ago
> monopoly
37.
by alpark3 3y ago
Obviously a sensationalist title, but a lot of details that seem true. I wonder if the open source models are going to go this route, imagine some sort of smaller, but still much more scaled/capable MoE model derived from the LLaMA mod
38.
by alpark3 3y ago
Repost?
39.
by alpark3 3y ago
I'm actually fine with the appointment setting, even if some others are bothered by the notion of an involved process. It's sort of like buying a toy anyways, so all the more fun for me to enjoy the steps of checking out. The thin
40.
by alpark3 3y ago
I just want a web client :(
41.
by alpark3 3y ago
>Developers wishing to continue using their fine-tuned models beyond January 4, 2024 will need to fine-tune replacements atop the new base GPT-3 models (ada-002, babbage-002, curie-002, davinci-002), or newer models (gpt-3.5-turbo, gpt-4
42.
by alpark3 3y ago
Honduras
43.
by alpark3 3y ago
I still think the largely-popular opinion that OpenAI is an evil corporation is wrong. It's easy to get caught up in the disagreements surrounding closed and RLHFed-to-death models, but I really do think OpenAI believes in the dangers
44.
by alpark3 3y ago
Is it reasonable to expect that some companies out there are training LLMs powered by Hyena instead of attention? Like a GPT-3 level size to bet on ultra-long context length?
45.
by alpark3 3y ago
Though I have premium, I actually don't mind Youtube ads that much. Compared to regular old banners ads throughout the internet, Youtube ads feel much nicer and interesting to watch. Much more production value, like TV ads(because they
46.
by alpark3 3y ago
But that number means nothing because it's not even close to the magnitude of energy we can begin to capture.
47.
by alpark3 3y ago
Speaking in terms of eye health, does it matter if the screen is right in front of my eye if the focal length is farther? Or is it still unhealthy?
48.
by alpark3 3y ago
I wonder how much of these changes are pushed by the local LLM shift we've seen recently. I would've expected them to totally focus on GPT-4 updates, but it's nice that we're getting 3.5 improvements. It's pretty cl
49.
by alpark3 3y ago
There's a bunch of LLaMA instruction-tuned variants on huggingface and new ones are coming out every day. I personally had the best experience with GPT4-x-vicuna: https://huggingface.co/NousResearch/gpt4-x-vicuna-1
50.
by alpark3 3y ago
I have access to GPT-4 through both ChatGPT Plus and the API, and at this point I only ever use the API through a discord bot. 25 messages/3 hours is way too low of throughput for me, and imo many people underestimate how powerful the
51.
by alpark3 3y ago
You're confusing Citadel and Citadel Securities. Citadel the hedge fund runs many different strategies, including commodities trading and quantitative high frequency strategies that are generally what people think of when looking at bo
52.
by alpark3 3y ago
There's a difference between taking a random() and then cubing it, and taking three different random()s and multiplying them together. The EV of the first is 1/4, whereas the EV of the second is 1/8. You can prove the EVs for
53.
by alpark3 3y ago
I heard something about the popularity of perpetual futures being owed to "insane leverage possible." Can anyone confirm/explain this?
54.
by alpark3 3y ago
I use GPT-4 through the API, but I can't help but hate the token/characterization pricing of these LLM APIs we've seen so far. Because the entire context needs to be fed back into the model, as my conversation gets longer, it
55.
by alpark3 3y ago
Stars boost visibility and is an important number motivating people who work on these for free
56.
by alpark3 3y ago
I use GPT-4 pretty consistently(set up a discord bot for myself). What I found myself doing was tending towards the most simple prompt that the LLM would still understand - If I asked a human expert the types of prompts I was giving GPT, I
57.
by alpark3 3y ago
Some variations. https://imgur.com/a/nJGcwMO I tried giving it more hints, but didn't seem to do anything useful.
58.
by alpark3 3y ago
A much better name than Mini-GPT4, in my opinion. Hopefully new cool projects that come out follow their example :)
59.
by alpark3 4y ago
Listening to the Lex Fridman podcast, Sam talks generally about how they want to hand off powers to the users. Users maybe not meaning the end-users, but the users of the API that create products on top of GPT, about how they tuned the mode
60.
by alpark3 4y ago
Me: Can you tell me about hackernews? Bard: ... Me: Cool. Who made it? Bard: ... Me: Tell me everything about Paul Graham. Bard: ... Me: what was my first question? Bard: I don't have access to that information. I am a large language m

More ›