shrimpx

“The Open in openAI means that [insert generic mission statement that applies to every business on the planet].”

To those who think this list will help them get into YC, or lament "why didn't I get into YC when my idea was squarely on this list":

The YC application is a sales pitch, and you're not selling your idea, you're primarily selling your charisma and capacity to spin vision and sell. Second, you're selling your chemistry with your cofounders and stability of your relationship. Third, you're selling your capacity to build, at least some usable prototype, but this a low bar.

At no point are you actually selling the concrete idea, unless you're doing something extremely specific that seems valuable and you're one of the few who can build it. For the rest, the idea is a rhetorical vehicle to sell the other things.

"FCC announces that artificial voices are indeed artificial."

How does the cabin pressure hold it in place passively? Isn't the pressure on the inside higher than the pressure on the outside?

`xclip-mode` looks like it should definitely be included by default. `cua-mode` is tougher because it messes with the default keybindings, making you type C-x twice (or Shift-C-x) for the large number of keybindings that start with C-x. That might be better for newcomers though, and bring more people to Emacs. Personally I would disable `cua-mode` if it were default.

Probably not. It looks like an autocomplete engine. But technically you can do that with an LLM, with a more complex interface. You could select a region and then input a prompt "rewrite this code in xyz way". And a yet more complex system to split the GPT output across files, etc.

You seem to be unaware of the mental and physical health toll these environments can take on people. Make money but become a burnt out, desiccated shell of a person.

I think they mean making people's lives miserable enough that they quit. Presumably they can live with whatever Amazon throws at them, and keep taking a paycheck until fired.

Anthropic made $200M in 2023 and projected to make $1B in 2024. That's a laggard <2 year old startup. I don't think LLMs are a fad.

Hey I'm Romanian, too. The latest iOS does what you want -- it has multi-language support and typing mixed English + Romanian is seamless now. Yeah it was a total pain to keep switching languages before iOS 17.

Yeah totally, I copy text from images all the time.

That's true, I have found that mildly annoying sometimes. But most of the time it's a win. It was really annoying manually switching modes over and over when typing in mixed-language, which I do fairly often. It'd be great if there was a setting though.

Yeah. Siri supports text input already, anyway. Siri is their ChatGPT-style bot that's going to keep improving.

Yup! TSMC wins either way.

- FaceID

- Facial recognition in Photos

- "Memories" in Photos

- iOS keyboard autocomplete using LLMs. I am bilingual and noticed in the latest iOS it now does multi-language autocomplete and you no longer have to manually switch languages.

- Event detection for Calendar

- Depth Fusion in the iOS camera app, using ML to take crisper photos

- Probably others...

The crazy thing is most/all of these run on the device.

Apple has been looking sleepy on LLMs, but they've been consistently evolving their hardware+software AI stack, without much glitzy advertising. I think they could blow away Microsoft/OpenAI and Google, if suddenly a new iOS release makes the OpenAI/Bard chatbox look laughably antiquated. They're also a threat to Nvidia, if a significant swath of AI usage switches over to Apple hardware. Arm and TSMC would stand to win.

Looks like clicking a reference adds the hash to the URL but doesn't scroll to the reference. If you load the hash URL directly in the browser you get a 404 page...

Since the article doesn't link to any example HTML article, here's a random link:

https://browse.arxiv.org/html/2312.12451v1

It's cool that it has a dark mode. Didn't see a toggle but renders in the system mode.

Overall will make arXiv a lot more accessible on mobile.

Yeah it seems like it's more about "blue bubble" status than it is about encryption.

On the iOS side people don't care, but apparently on the Android side they do?

It's because SMS is unencrypted, and this is an effort to have secure communication between iOS users and Android users (since iMessage is Apple-device-only). There's also a status factor, where Android users are lowly "green bubbles" in iMessage, which signals decreased security and questionable support for the interactions available to iMessage users.

Of course everyone can switch to Signal and this is a moot point, but millions of Apple users use iMessage and aren't likely to switch.

The microphone trust aspect in the article seems like a red herring distracting from what could be a clearer point.

Facebook literally takes your data from their apps and internet, tracks your behavior on the internet, and feeds this data into models of you. These models are so accurate they can sometimes basically predict what you're thinking. Hence the layman jumping to the conclusion that they must be spying though the mic.

A LLM company like OpenAI, and their partners, employs almost literally this exact model. Grabs data from whatever sources to improve their models, to increase the likelihood you'll keep clicking where they want you to click, to monetize you.

Basically ideas are a dime a dozen.

There's a spectrum of ideas, from groundbreaking to "dime a dozen". In tech startups, and in almost all of computer science, most ideas are a dime a dozen, and the value is in the execution.

But clearly, some ideas are groundbreaking. Einstein rightfully gets the credit for an on-paper hypothesis that wasn't proved until decades later via a chain of critical discoveries and experimental innovations by other people. It's legit to call it Einstin's relativity, and not Mossbauer/Hay's relativity.

Btw Andrej Karpathy builds a Shakespeare LLM here:

https://www.youtube.com/watch?v=kCc8FmEb1nY

Which provider, if you don’t mind me asking?

I agree with that.

IMO employers discriminate a lot, for random off the cuff reasons. The largest, most lawyered up companies that are glaring lawsuit targets have more of a process. But small-to-mid companies discriminate like crazy. It largely comes down to some guy with authority giving a thumbs down, for whatever reason he wanted. "I didn't think he was that good at SQL" may very well be "I didn't want to hire another Indian" or "This person worked for that guy I met, who I think sucks, which is like a bad vibe". Not hard to imagine having porn lawsuits on your record disqualifying you from a bunch of interviews, but the reasons stated being bullshit technical/cultural fit reasons.

That’s kind of cool. Also related to local-first, maybe.

Thanks I updated to clarify.

I guess it's kind of like doing all the things a join query would do, like mapping rows together, filtering, sorting, and grouping, but the final output is denormalized into several "tables" without duplication. From what I gather, the main "win" is the size reduction in the result set, which can impact network and memory usage. Yes you can do this yourself by running multiple queries, one on each table. Presumably by paying the latency of multiple round trips.