llmfan

OK, I'll ask.

DeepSeek R1 gets it right:

The species formerly known as Brachiosaurus brancai was reassigned to its own genus, *Giraffatitan*, through a series of taxonomic revisions. The key steps were:

1. *Initial Proposal (1988):* Gregory S. Paul first suggested separating B. brancai from Brachiosaurus in 1988, proposing the informal name "Giraffatitan" as a subgenus. However, this was not widely accepted at the time.

2. *Formal Reclassification (2009):* Paleontologist *Michael P. Taylor* conducted a detailed comparative study of Brachiosaurus specimens. In his 2009 paper, he formally elevated B. brancai to the full genus *Giraffatitan brancai*, citing significant anatomical differences (e.g., limb proportions, vertebral structures) from the North American Brachiosaurus altithorax.

Thus, *Michael P. Taylor* is credited with the formal reassignment of the species to the genus Giraffatitan in *2009*.

actually, a more practical idea would just be a generic "pandemic risk" site.

Seems too narrow to be a website.

Maybe a site where you get overviews like this... to a prompt.

"Make me a dashboard on the current Moon race. And one on Taiwan invsaion. And one on AGI timelines."

"Yes, send me an email if there's a major shift in any of them"

If an AI agent understands you (and book writing, and the topic of the book) well enough then it should be able to write you a pretty nice bespoke book.

I do suspect that interactive media is just strictly better in theory. But maybe there will be a period of time where bespoke AI-generated books make sense.

I mean, obviously it's a numbers game. There are only so many hackers, but there are a lot of Ivy League people and FANG people. So YC chose to open their gates. Whether this will be a profitable decision is an open question.

I can tell just from looking at most recent YC founders that they're not hackers.

I could be wrong of course. I'm probably wrong. Maybe these people who are attractive, care about popular fashion and becoming popular on tweeter are actually just as capable hackers as all the hackers I actually know (none of whom care about those things). I admit YC probably knows better than me.

But if I'm not wrong, and these people really aren't hackers, it means YC has decided that normies > hackers when it comes to YC. I really don't predict that's going to end well for YC.

LLMs are already better than books for exploring some ideas. But in conversation form.

Until we get better versions of o1 that can generate insights over days and then communicate them in book form the loss of interactivity and personalisation makes LLM books pointless.

Doesn't Microsoft have a decent amount of ownership in some aspect of OpenAI's labyrinthine structure?

Take Anthropic for a more legible example: Amazon and Google both gladly bought large stakes in it.

I think this is one of those "controversial" topics where we're meant to be particularly careful to make substantial comments.

You can judge them to be not a generally super-trustworthy brand.

But I would grant them that their responsibility for the deaths of these people is limited.

https://hub.scroll.pub/onlybreck/

That was the intention behind my comment! No need to thank me tho.

hnews tries to say one positive thing challenge impossible

i always love any idea for curating high iq internet community

yeah was disappointing

intersection between AI and sex and relationships is very interesting and important

Current models are very cheap. Like GPT-4 cost ~$40 million to train, right? Once we get 100 billion dollar models I suppose it'll be different.

I mean, consistent mediocre releases is exactly what we have gotten out of OpenAI.

But we know they started training Orion in ~May. We know it takes months to train a frontier model. Lack of release isn't promising or worrying, it's just what one should expect. What is promising is the leaks about the high-quality synthetic data that Orion is training on. And the fact that OpenAI seems to be ahead of all the other labs which are only just now beginning training runs on next-gen models. OpenAI seems to have a lead on compute and on algorithmic innovation. A promising combination if there ever was one.

"just a regular LLM except [trained on very different data]."

I'm not saying there's some big moat, anyone can read https://arxiv.org/pdf/2305.20050, but not all synthetic data is created equal. Strawberry I'm sure generates beautiful, valid chain-of-thought reasoning data. Wouldn't surprise me if OpenAI is just significantly ahead of the competition.

Doesn't that just depend on what LLM you happen to be using? And doesn't Perplexity support many different LLMs?

An option to side-by-side comparison could be cool

So many different areas of knowledge can be leveraged, as long as you're able to experiment and learn.

It should be called prompt science.

The article is literally presenting the case that crypto is ALL a scam (emphasis not mine). Obviously not true, and attempt at logical argumentation in article is painful.

The moat is compute and Elon has enough money and connections to jump the queue with providers

Everyone knows you have a secret password.

Not so many people know that YC has a secret [insert whatever it is that they have that is secret but that we don't know about here].

I'm seriously confused. What's wrong with it? What are the complaints about? It's some of my favourite content, and seems like it couldn't be more on-topic and helpful.

Notes:

- You can't CTRL-Z back to before you clicked the "save" button

- I couldn't figure out how to do links properly within a table

http://sand.scroll.pub/anime/

Your website lived up to the title's promise, congrats.

I had to click the download button.