weiran

Only when my usage runs out

25% more usage sounds about right given the other token costs are down about 20%? I don't think cache read is a big portion of the overall cost.

Not that I've seen. The only follow up I've seen from someone was it only happened if you had a free account and not paid (which would explain why I'm unaffected)

I've been using ZCode since it's initial release and can't find any of this in my data. There aren't any logs showing capture or upload, and I don't even have a ~/.zcode/v2/checkpoints/ directory.

So unless they've cleared it all with a recent update then it doesn't seem to affect everyone.

Pi is a very basic harness by design. On the other hand OMP is a bloated mess of other people’s workflows.

The trick with pi is to extend it yourself as you use it. It’s pretty easy to do.

You can get around the checks with careful prompting. You’re just interested and want to learn more about it.

Try posting anything even mildly positive about AI on Bluesky and you’ll experience instant hate.

You need a paid dev account for that. I think self signed installs can run for a year.

Also the part people forget is the significant coordination cost of having multiple teams work on multiple native codebases.

I hate Electron apps with a passion, but I can understand why the AI companies are using it.

There’s nothing surprising about Bluesky’s audience being toxic to any mention of AI. It’s the dogma over there and anyone who doesn’t agree has long left or never joined.

Yep these software benches are only good at testing how well they can one shot. For the kind of attended/assisted development most of us do with agents it’s hard to find a benchmark that reflects my own experience of the frontier models still being quite far ahead.

It’s how sandboxed Mac apps work today.

> What’s wrong with people?

Confirmation bias.

Reasoning levels are often just injected system prompts so not a great way to fingerprint models.

PWAs hold themselves back, they don’t need any help from Safari.

TBH I don’t think I’ve seen a lidded chip on anything other than a CPU.

The self hosting community is incredibly anti-AI for some reason. I’ve never understood why as using agents has made managing and maintaining my home lab system so much easier.

After trying it out for a bit, it feels like a slow GLM-5.3 but with vision.

The only way the Chinese models financially make sense is if you use their subs or host it yourself. DSV4 is especially token hungry.

You should be able to run Q3 and maybe even Q4 quants with 32GB. Even with the GPU as you can up the max RAM allocation with: 'sudo sysctl iogpu.wired_limit_mb=12345'

It figures that the thermal pads are applied by hand, when I replaced mine they weren't very well installed in the first place.

You'd like to think that a GPU like the 3090FE would go through QC that include some sort of stress test. It would only take seconds to see the VRAM temps would be well out of spec.

Unfortunately Web API 2 doesn't seem to work on it.

I get around $30 a month from a Hacker News app I made for iOS.

Unlikely to have a major upgrade until the Broadwell is out in mid-2015.

Lack of quad core is the only serious issue I have with the 13" now. Hopefully Broadwell will let Apple put a quad core into the 13".

The new Thunderbolt version which works like a native port at full gigabit speed.

Try this bookmarklet: https://github.com/joelcarranza/particular-pinboard

How is Apple magically supposed to know that you've switched phones?