SparkyMcUnicorn

I think the biggest selling point of a codex sub vs a claude sub is that you can use codex subs in any harness you want.

> You can also use Tailscale serve to get a HTTPS url like Cloudflared but it's only visible to your tailnet

For public access, similar to these cloudflare tunnels, there's Tailscale Funnel.

https://tailscale.com/docs/features/tailscale-funnel

It's almost comical how annoying it is. I just want to try it.

I still think the only reason to use jQuery is if you need IE11 support.

Otherwise, it's really simple to do all the things you mentioned without the complexity/heaviness of jQuery. Something like this:

``` const $ = s => { const r = {}; const els = [...document.querySelectorAll(s)]; r.text = t => (els.forEach(e => e.textContent = t), r); r.on = (ev, fn) => (els.forEach(e => e.addEventListener(ev, fn)), r); // ... return r; }; ```

I'll watch Marques when I want to get caught up on stuff like this, because he usually has a minimal fluff video on the subject.

For foldable phones, this might be a good pick to get caught up (5 months old, but minimal crease): https://www.youtube.com/watch?v=eFUB_jL_XcM

Closed source, solves a different problem, and just looks like generic agent sandboxes/VMs?

What am I missing.

Have you given Ornith-1.5-35B a shot?

It's been a pretty decent step up for me compared to Qwen3.6

https://news.ycombinator.com/item?id=49362401

I run into this several times per week.

Sometimes the context where Fable drops to Opus 4.8 is enough for Opus 4.8 to do a first pass on everything, then I start a new session and Fable will happily refactor/improve it.

Super annoying.

If active weight bandwidth is everything, then wouldn't Qwen 35b a3b (FP8) be roughly half the cost of dsv4f (FP4)?

dsv4f appears bigger and more memory/computationally expensive to run any way I look at it.

Yeah, I would appreciate if someone could make sense of the pricing differences between these models. How can a provider run DSv4F at lower cost than a 27B dense or 35B A3B model?

Does it come down to utilization and/or specific model tricks and efficiencies (attention, kv cache, etc.)?

DeepInfra prices:

Qwen 3.6 27B: $0.32 in / $3.20 out

Gemma 3 27B: $0.08 in / $0.16 out

DeepSeek V4 Flash 0731: $0.08 in / $0.18 out

Qwen 3.6 35B A3B: $0.10 in / $0.95 out

https://openrouter.ai/qwen/qwen3.6-27b

https://openrouter.ai/google/gemma-3-27b-it

https://openrouter.ai/qwen/qwen3.6-35b-a3b

https://openrouter.ai/deepseek/deepseek-v4-flash-0731

16 MHz is the chip's clock speed. Arduino PID is more like 100 Hz to 1 kHz.

Human deliberate movement runs around 2-10 Hz I think.

This source appears to have more images and info: https://www.frontiersin.org/journals/marine-science/articles...

Relevant: https://blog.comma.ai/011release/

If you use Claude Code, there are plenty of projects/tools that let you see what your usage would have cost at API rates, then you can cross-reference that to your "usage" in Claude. I run through enough token in a few days for the subscription to pay for itself.

But I agree that advertised visibility into this would be nice, so we could compare the different providers up-front.

I really like the DM Pro and have it deployed to an office of about 50 people. It's a pretty no-fuss solution and fairly simple to manage.

For my personal setup, I decided to go with OPNSense and I couldn't be happier. Much more control, at the cost of being a little more hands on.

I think the best (rough) comparison here is MacOS vs Linux (or more accurately in this case, FreeBSD).

Are you suggesting that Chrome is going to allow individual websites to override the browser DNS settings? And also skip bootstrapping?

I'm very confused by this comment.

Freedom of DNS choice has nothing to do with DoH.

And you can definitely mitm a non-configurable DoH resolver if you absolutely needed to do that, as long as you can add your own trusted CA on a device.

How does DoH remove any capabilities of what the resolver can respond to queries with? I block ads via a DoH resolver.

I don't think we get to blame these issues on "vibe coding", they've been around for too long.

Here are some llama.cpp benchmarks for it: https://www.phoronix.com/review/intel-arc-pro-b70-linux/3

This comment is against the guidelines: https://news.ycombinator.com/newsfaq.html

I don't want to accuse you of something without proof, but the 3 other comments here are from brand new accounts and all smell very strongly of self-promo (and also appear to have used an LLM to generate their recent activity).

temperature, top_k, and top_p don't exist on Opus 4.7 (or 4.6?).

Related: https://xcancel.com/bcherny/status/2044831910388695325#m

https://platform.claude.com/docs/en/api/messages/create#crea...

They should have kept an eye on the prediction markets.

On every keyboard I get, I swap the switches out for silent tactiles[0][1] that I've selected through trial and error. Quiet is really nice (my mouse clicks are louder), but the way they feel is fantastic.

I wish there was a brick and mortar that let you try out a good range of these switches. Places like microcenter have the popular standard choices, but there's so many other switches out there that are just worlds different.

[0] standard preference: https://a.co/d/03j6Boy0

[1] low profile preference: https://a.co/d/06yVB6jg

Opus[1m] has heen the default model for max subscriptions since 2.1.75.

https://github.com/anthropics/claude-code/commit/48b1c6c0ba0...

Explained in the methodology at the bottom of this page: https://marginlab.ai/trackers/claude-code/

I was implementing OSM the other day (with attribution of course) and noticed Carto seemed to be one of the only players with good dark tiles.

Seems like we can't use them for free, even with attribution, unless I get a grant?

I've found my research on certain topics like this becoming less reliable these days, compared to just trying it out to form an opinion.

Your repo says it's open source, but it's missing the source.

https://github.com/ctxrs/ctx