it's mostly about AI training at this point. the software for this only supports CUDA well.
HN user
StochasticLi
don't illegal immigrants commit less crimes per capita?
Exactly. Also, it's incredibly affordable, especially when you buy it on Black Friday.
I can see room for improvement about the UI, but I like that it works everywhere and I use it to, obviously fix grammar and mistakes, to add/change our company's specific vernacular (finance), and to change common LLM sentence forms so that the readers don't notice that content was made by AI. Ya know, all these em dashes, "testaments to", "this isn't just X, it's Y", etc.
My favorite function of the UI is that, on any website, pressing TAB just fixes/improves what I just wrote before.
I'm sure they haven't created their own model. Probably fine-tuned it + constantly updating a database of the most occurring phrases from all LLMs.
That's where the research is going.
Grammarly has a function that replaces phrases that are typical for LLM output. Good luck coding that in a few hours.
I'm paying for most AI models top tiers and Grammarly. Grammarly is a phenomenal tool. It's not that LLMs can't do it. Well, they can't, but the more important thing is Grammarly's UI.
I think we can reasonably expect they will become non-stateless in the next few years.
I don't think it matters. Here is why I would pay 10x more for using Apple devices than other devices:
1. The ecosystem is just much better, ie. the quality of apps is in another league compared to apps on Android or Windows.
2. Many amazing apps that are at the core of my company's workflow simply don't exist on other platforms at all. With no comparable equivalents.
3. We get many apps earlier. Even up to a year earlier. Example: Clubhouse. Running a company, my marketing team can't afford to be late to a new social media platform, risking that it becomes the new status quo. I can name a dozen apps that were 6 to 12 months late to Android vs. IOS :(
If these change I would happily move to Linux, or Windows (once they make the OS usable).
We use Linux servers for running Python/Node Scripts, and APIs.
We use 1 Windows machine to use a very specific data science app that only exists for Windows.
Before anyone says I'm a fanboy. I'm here to generate revenue for my company. I don't give 2 sh*ts about AAPL.
---
But this is just me.
Why it won't matter to the public: The kids don't care. The want an iPhone.
The CUDA ecosystem is their moat.
Jacob & Co. Watches recently
I would like to know what inference speeds they are achieving exactly on what hardware. I skimmed and searched the article and didn't find that info.
What do you think is a proper solution? 30% being pro genocide is insane.
Please name them.
I'm running a company so it's more than a project.
No. I worked at Meta, and these are mostly mid-skilled engineers who want to make money while working a few hours a week and pretending to work the rest. They would never attract top talent because top engineers can earn 10 times their salary on their own without leaving the terminal nowadays. Even if they offered $10M/year, I feel these kinds of people can't be bought.
It's just a job. You will get fired if you question people above you.
For a similar reason I kill all my unpinned tabs automatically every 24 hours. If I didn't do/read it, it means there are more important issues.
You don't get withdrawals from cheese, and the cure for these withdrawals isn't Suboxone
Johns Hopkins: "Kratom (Mitragyna speciosa): Pharmacology and use of a naturally occurring atypical opioid."
Because it's an opioid?
They give me massive anxiety.
read some books on parenting
Can you explain how I will meet people at the library?
In 1 sentence: Do the opposite of trying to solve the Collatz conjecture.
No. You're paying for a minute of audio, which will be more packed with speech, not for how long it's being computed.
This whole thread is a great example of the developer vs. user convenience trade-off.
A single, pre-compiled binary is convenient for the user's first install only.
Because rich people can reduce their income to 0 or close to 0 via different legal ways like LLCs, funneling through other corps, certain investments(which are basically tax-money laundering schemes), being tax residents of tax havens, etc. It depends on the country what you can do. In Germany, it's basically 0.
Berkshire Hathaway can't change this.