notfried

True. I was reacting to the article mentioning that April was when the usage exceeded salaries. Mythos was released in restricted preview through Glasswing in early April, and I presume was available internally for at least a few weeks before that, which helped tip the numbers.

It's kind of expected, no? If I was an Anthropic employee expected to dog-food the product and be on the frontier of usage, I'd probably be using Mythos all the time at the highest thinking level. And if I am not being metered, I'd probably do more stuff that I know are likely low yield since I am not worried about a quota. This would probably take my "API-equivalent" usage beyond the average salary in far less than a month.

This was quite ironic:

"I mean, there doesn't necessarily have to be more."

"A lot of people are focused on like taking over the world or doing like the biggest thing, getting the most users."

Then he adds: "Part of making a difference and doing something cool is focusing intensely."

That bureaucracy is there for a reason, you often need bureaucracy to serve bureaucracy. About 85% of Microsoft's revenue comes from B2B/Enterprise/Governments. From their latest fiscal year report [1], the top 4 revenue groups were:

Server products & cloud services (39% of revenue) M365 Commercial products & cloud services, excludes Consumer (30.7%) Xbox / Gaming (6.6%) LinkedIn (6%)

[1] https://app.edgar.tools/companies/MSFT/disclosures/segments

Meta will surely appeal, and will likely get the fine reduced to like 10% of that. But for comparison, in the recent multistate settlement over social media addiction by 47 states for $17 billion, there was a provision in that settlement to get $459 million [1] in exchange for releasing Meta from any state claims related to Cambridge Analytica. So those 47 states got $459 million for what New Mexico will likely get a few billion for.

[1] https://www.techpolicy.press/the-459-million-cambridge-analy...

MyFonts (I'm not affiliated to it) has a visual search for fonts, and searching for this design yields 94 results that are not identical but collectively share the same patterns. Now, fonts are horrible when it comes to licensing. Like, in this case, a font I picked has a 2M page view limit per month and you pay $660/year for that highest tier, so I get why a company would find a custom font a reasonable decision, visually and/or financially. But 4.5 years is absurd.

It makes sense they'd eventually do that. The Apple TV didn't launch with an App Store, and it took 8 years after its first generation to get an App Store! I'd guess Apple sees that most apps wouldn't make sense for the HomeView, and they'll likely restrict what categories of apps can launch on it.

No one with real money or power wants to cross OpenAI or Anthropic just yet. It's a closed ecosystem. OpenAI hacks HuggingFace, Nvidia soon after buys HuggingFace while it also before backed OpenAI. In the end, they just handle it all behind closed doors like family business.

I saw the landing page and I was like, "no way, Claude built that in real-time? Incredible." And then it loaded and it looked like a videogame from the 1990s. Pretty impressive for AI to have built it , and for something like that to run in the browser, but it's so clunky and ugly that I couldn't try it for more than a few seconds.

I hate it when products promise something so breathtaking in posters or trailers that is utterly detached from how the products looks or works. I might have tried it longer had that landing page not set this high of an expectation.

Overselling a product may getting you a higher conversion rate up to a point, but it will only hurt you and crater the rest of the funnel.

They also said this in the email announcement I got:

> As a thank you for being an existing subscriber and to help you continue to build your most ambitious ideas and transition, we’re also giving you a one-time grant of 62,500 usage credits (a $2,500 USD value). You’ll see these appear in Settings > Usage. They will expire on December 31, 2026, and use of the credits is governed by the Service Credit Terms.

I have no idea how much 62,500 usage credits are. Given the value they suggest I feel it's like one extra week of usage.

5.3 was incredibly better than 5.4/5.5. I stuck with it for months after 5.4 was released, and kept testing 5.4/5.5 every now and then but they both were too inconsistent, too rash. I switched to 5.5 a few weeks ago and now regret it, but I am no longer seeing 5.3 as an option to use, only 5.3-Spark, which is trash compared to 5.5.

When an architecture company seemingly uses AI to render mockups, they really need to ensure consistency and accuracy. It's not that difficult nowadays. It was quite confusing trying to understand the differences in design between pictures and to compute why the tunnel seems so short compared to the mountain, until I realized it must have been laziness; not laziness because they are using AI, but laziness to do their job right.

I was surprised to see Apple exposing the device boot time. Not sure why a typical app would need this, and in a world in which iPhones are infrequently rebooted, the combination of this with other fields must indeed be very deterministic.

Why is the author considering Claude Code a "real developer workflow"? Unless you're doing complex tool calling, is CC really resource-heavy?

If AGI were to happen, or if AI became a trillions-of-dollars-generating industry, you wouldn't want to have your data-centers which might be the most valuable thing on Earth be located in a foreign country. All this investment in infrastructure is not purely based on where the industry is now, but predicated by where those who are bullish about it think it will be in 5-10 years.