wahnfrieden

It’s wrong

I’m seeing hundreds of review loops before settling, with Astra/Sol

Lonsdale was also accused of rape and is quoted as saying that women need to be raped to learn to respect men.

The info you’re citing is for API not Codex

You’re sharing outdated info

And comments complaining about other comments too

Liquid Glass was rushed out because their flagship WWDC feature, Apple Intelligence, wasn’t ready as expected. They knew they didn’t finish refining Glass as they had intended to. They had unfinished work with Glass that got pushed to this year due to the last minute change in direction. The refinements we got weren’t only the product of feedback and they already intended to revisit the design.

This situation is a bit different where they’ve taken several years delay now to refine their AI strategy before releasing it.

If they let users permanently opt out of their AI rollouts this early on, their stock will plummet. They’re going to focus on making it better to drive adoption and value instead of letting people remove themselves from exposure to what they have in the pipeline.

Time to permanently offboard from Apple ecosystem then

The Buchla Easel is semimodular and has preset cards and a memory recall card

The Easel has internal wiring that minimizes excessive wiring needs

I ordered the Buchla easel briefcase

It’s a lot of work

Apple does many crazy custom things in their own apps. Their advice is not sacrosanct. It just means you must prepare for less notice than they had if you want to compete with the frontier UI aspects that Apple apps achieve.

No it’s not

That’s not rehosting

Edit: iPhone autocorrected my OP which meant to say rehosting not reposting

It’s the rehosting of the content not the scraping

Edit: not a moral stance

Salomon has been a popular choice for everyday use for at least a decade. They’ve even been collaborating with high fashion brands for over a decade, like Boris Bidjan Saberi, Commes des Garcon, Margiela, Takahiro Miyashita etc etc with twice-yearly releases over years and not random one-off limited drops.

This has led to fashion boutiques carrying the non-collab Salomons as the demand has diversified. So people are exposed to the brand and have access to it at a variety of kinds of shops, not only at running outlets.

Well, those training inputs reflect how human thought and action are documented or otherwise expressed on paper. Humans have behaviors and mechanisms that these expressions don't translate.

The X3 seems nicest as it’s smallest and a good deal lighter.

Ofc

And no Android

How much time have you spent in Japan? The bare foot never touches the genkan floor. It's alarming to do so.

Since you can't imagine it, you could consider looking up some videos of their use before mocking what you consider a fantastical impossibility. In fact it's so easy that children and the elderly succeed at it.

Wrong

No

It's highly dependent on workload

If Astra is per benchmarks so much more token efficient than Sol, why did they limit its use in ChatGPT to ~16% as many messages compare to Sol? When Sol is 40% the price of Astra, why do they give Sol Pro (in ChatGPT) 6x as many tasks?

If it were strictly true that token efficiency makes Astra cost around the same per task as Sol then there'd be no need to limit it to 16% as much access.

It's because per-task token use is highly variable, not as universally true as you claim

I gave more details here https://news.ycombinator.com/item?id=49559593

Besides the usual tricks to optimize token efficiency, token use can be highly workload-dependent.

Terra is not enough for many tasks - even Sol is not enough. That's why I'm eager for Astra.

I'm doing a lot of rearchitecting/refactoring and hardening in a large handwritten codebase; iterative performance and storage optimization (some areas I've been iterating on since January, with tremendous new progress unlocked by each new model release); offline-friendly, multi-device realtime sync with complicated requirements; and various natural language processing and other such problems that are essentially unsolvable but can become more accurate and better tested for accuracy through iterative work. Off the top of my head.

Some of these tasks involve a lot of code reading or other inputs, or reevaluating work. Token efficiency is no help there, if the input can't simply be skipped. Cheaper models are sometimes bad at summarizing or highlighting the right parts, depending on the task.

I also don't use subagents except for Luna. I'm mindful of cached sessions and start new ones often to avoid loading in full contexts (often with some kind of handoff doc or skill).

The only subagents I use are Luna.

My use cases are iterative and sometimes require reading a lot of code or reevaluating work.

Token efficiency is near meaningless when the workload is input-heavy. It can't always just choose to read less, depending on the task.

I can have cheaper agents do the reading but it's not appropriate for all use cases because they'll misjudge and choose the wrong things to emphasize, summarize, extract for the bigger model.

Many threads.

I use new threads if relevant old one is uncached. (Often using a skill or doc for handoff instead of requiring full context gathering again.)

I get involved in architecture and specific implementation direction. The codebase is 8 years old and mostly handwritten.

Mostly coding. Some QA.

No cron/CI agents.