maeil

32 ms·

631.
by maeil 2y ago
The article says the following: > If buses arrive exactly every ten minutes, it's true that your average wait time will be half that interval: 5 minutes. Which means that "not arriving exactly on schedule" is indeed a requ
632.
by maeil 2y ago
Harrowing read. I've believed this for years - if there's a single text that every single high school student in the world should be forced to read, it's a text on this subject. The effects of nuclear war. The close calls. Th
633.
by maeil 2y ago
One of the things I love most about HN - you think of a hypothesis, and right after you see someone else who has already posted it. Very curious about this one as well. Though I don't imagine it being the "actual cause", just
634.
by maeil 2y ago
You intentionally moved to that area, and it even sounds like you own real estate there. Yes, it will impact your lifestyle, and you may have to be more careful about what money you spend, depending on how you're currently spending it.
635.
by maeil 2y ago
> I'm constantly surprised how many well-off people try to pass themselves off as "one of the poors". The amount of people I've seen on here saying "It's just not an option to quit Google/Meta no matter
636.
by maeil 2y ago
Genuinely curious as you have experience with fine-tuning, which I don't yet. Given how simple it has become to get the correct response shape, do you feel there's still a point to it? If you need JSON and want to be very sure you
637.
by maeil 2y ago
It's not about population vs. individual. The underlying principle is the assumption that we (the people taking the bus) arrive at a random time. And we're more likely to arrive in a shit interval (because they're longer) tha
638.
by maeil 2y ago
This was easily the most memorable thing I learned during my statistics degree! Nothing else has stuck with me this well.
639.
by maeil 2y ago
A few weeks ago I had a eureka moment to describe it: GPT writes just like a non-native speaker who has spent the last month at a cram school purely aimed at acing the writing part of the TOEFL/IELTS test to study abroad. There, they a
640.
by maeil 2y ago
You make good points, but to be fair, I feel it's more the beating around the bush that people take issue with.
641.
by maeil 2y ago
On the contrary, the comment history makes it very clear. Pages and pages of relatively short comments, not a single one written in a remotely LLM-reminiscent style. Then, within a very short period, multiple very long comments in exactly t
642.
by maeil 2y ago
It is, and I'm curious what dang and HN's plan is wrt this issue going forward. On one hand, the "assume good faith" has been a core tenet of this community. At the same time, LLM-generated walls of text aren't good
643.
by maeil 2y ago
The former would seem ineffective, surely anyone buying such a dataset would just remove the "+___" part in emails that have them?
644.
by maeil 2y ago
While of course this being a possibly valid legal argument is completely inane, I must assume Disney is doing this for precedent, not for this one case. They want to be able to use this absurd line of reasoning in any future lawsuit and are
645.
by maeil 2y ago
So Disney is "arguing that he agreed to settle any disputes with the entertainment giant and any of its affiliates out of court when he signed up for a free trial of its streaming service Disney+" for a case about "the death
646.
by maeil 2y ago
> It's an interesting hypothesis, but it's entirely possible that separated from Google, Chrome is either not a viable business model (browsers don't make money on their own) No, that's not possible. Google pays $20 b
647.
by maeil 2y ago
Going to add my personal favorite: https://artificialanalysis.ai/models/llama-3-instruct-70b/pr... What sets them apart is that they have speed and latency as well.
648.
by maeil 2y ago
I guess for a wizard it does make sense for a spell-gone-wrong to have chopped off two of his fingers.
649.
by maeil 2y ago
The models are still of a level where for less common/benchmarked tasks, there's often only one model that's very good at it, and whichever is 2nd best is markedly worse, possibly to a degree where it's unusable for anyt
650.
by maeil 2y ago
Not really. There's a hundred benchmarks, but all of them suffer from the same issues. They're rated by other LLMs, and the tasks are often too simple and similar to each other. The hope is that just gathering enough of these benc
651.
by maeil 2y ago
Your concerns are valid. Two more things concerning Chatbot Arena: - The prompts people use on have an incredible sample bias towards certain tasks and styles, and as such are unrepresentative of "overall performance" which is wha
652.
by maeil 2y ago
Their lawyers may well share their concerns, but in the case of X, those lawyers may simply be getting ignored. This isn't a normal company.
653.
by maeil 2y ago
There's a reason why the world outside of the US and Eritrea sees citizenship-based taxation as a completely bonkers concept.
654.
by maeil 2y ago
You wouldn't, because the shareholders would absolutely revolt. Europe is responsible for >20% of Apple's revenue.
655.
by maeil 2y ago
> “I think there’s a real social media fatalism that has set in, that it’s just irredeemably toxic and never going to get any better,” Pariser said. “The goal here is to demonstrate that local conversations don’t have to be toxic. That’s
656.
by maeil 2y ago
We mostly do multimodal tasks (vision + text), and there the differences between flagship models are still much bigger. For us, the benchmarks showing all of them being close are pretty meaningless, it really depends on the task when vision
657.
by maeil 2y ago
Right. Another commenter mentioned "a nice sit down restaurant", and some of the other replies imply they're just not used to not drinking without a straw at all, hence my confusion. Now fine dining is on the extreme end, but
658.
by maeil 2y ago
Then do you have a suggested alternative?
659.
by maeil 2y ago
Are you replacing LangGraph with something else? I've luckily avoided LangChain exactly because it's a black box, but been on the lookout for some kind of DAG visualization of LLM flows. LangGraph pops up first on any related sear
660.
by maeil 2y ago
Same, broke product. Issue is that there is an (albeit small) subset of tasks where Claude is first in class and can't be replaced by any other model without a decrease in quality. Even then, taking away the same lesson here, and leadi

More ›