batmenace

Listen, I'm not saying you're wrong, but I will say I hope I'll at least get to update this comment down the line while holding that gorgeous looking phone in my hand

Maybe the Copilot infra for OpenAI and Claude is via Azure?

So basically blot.im?

Couldn't agree more. Spent the evening reading though some of their work, and it's all so... fun. Just exploring fun and potentially useful ideas not restricted to what tools already exist (which seems to be the case much of the time)

I think it’s a neat idea (maybe because I was talking over ow to do something similar in Python, specifically for financial analysis). Definitely feels like an early stage, but I would love to see where it goes.

A part of me likes this. Every now and then, isn’t it nice to pursue something and just see where it goes, without a monetization strategy or similar?

Feels deliberate, just like ‘das sequencer’

Well the thinking by google has surely been that Apple Search (iSearch?) would be competitive if Apple percieved it would be in their interest to build it / work with another partner. It's not impossible that someone else would build a good / better tool. Since all these tools also significantly improve with use and feedback, switching the default would probably have a massive impact

Yes, humans operate similarly. We gain knowledge through learning and assimilation of experiences. But there is a cost associated with each new input, in one form or another. You can read and reproduce the NYT articles or their content and build based on what you have learned. But aside from piracy, you have to somehow pay a cost associated with access (ads, subscriptions etc). The LLM does not do so in an equal way.

Well if the company goes bankrupt, theoretically other companies can buy e.g. 3M's patents / processes / subsidiaries and continue their production. THe company isn't unprofitable, so it would be sold off in parts and the proceeds of the sale (theoretically) would go towards settling lawsuits.

> They didn't.

How would you know? How would we know without actually auditing their books?

IPO doesn't have to mean ads, though. There are other ways to monetize a user base and to generate money based off of a 'community'. They're just ways that require more time, more concerted effort and planning to execute.

Also presumably the reason for making the first one a 'pro' version -- and beginning the presentation with a focus on work. Easier pitch that way, and once it can establish itself / there is an ecosystem, you can probably get more 'consumer' focussed versions with less compute power or that are plugin only etc

Doesn't really answer what they all do... What are they up to day-to-day

Your comment supposes that it has to be extreme -- rapid growth or stagnation. Why do you think a responsible CEO / management team cannot plan for a 'moderate growth' pathway that sees steadier levels of hiring and firing more akin to what the rest of the industry shows?

I wonder if for tech companies part of the answer is also that moving people around is an 'easy' way to play with numbers. You go to the market / to your investors and say ''Hey, we've hired 1,000 people over the last 12 months, we must be doing so well [give us money]'', and likewise ''We're gonna lay off 1,000 people soon so we're all good here [give us money]''

Colours of the company that's been developing it, presumably the final version will be nasa / white color scheme

Well that's far from the complete picture. It's not that the gov't is backstopping all bank stupidity -- the new facility simply says that for redemptions, US government bonds and MBS can be valued at face value not market value. Only for the purpose of ensuring liquidity for redemptions

Fermat's Library offers something that comes relatively close https://fermatslibrary.com/margins