rbinv

I think the question meant "to what extent did they mess it up" not "how did it happen".

32 GB GDDR 7

I used Wireless Diagnostics.app on macOS ("Scan" window) and looked for the 2.4 GHz radio's MAC address which was nowhere to be found. But yeah, you have to have a hunch.

I've had similar issues with clients (Blink camera sync modules) connecting to far-away APs instead of the much closer router. I could not force them to connect directly to the router no matter what.

Turns out the 2.4 GHz radio of my router had died (Fritzbox 7590, apparently a common issue), thus forcing 2.4 GHz-only clients to use those other APs.

Well... has it been the best so far?

Thanks for the meme chuckle. For the uninitiated: https://i.imgflip.com/ayibmd.jpg

Ente feels a lot like Smugmug did back then.

Great thinking on your end! I will run the smoke-test once you're ready.

I agree. Mowing the lawn is repetitive in a good way, almost meditative and flow-state-like.

What city are you talking about specifically?

People will pay for a good workflow/orchestration layer around coding models, see Cursor. I agree about the moat though, seems fragile.

I find it refreshing to see dev tools competing on latency again. Has anyone tried both Conductor and Zed for agentic coding on a real project? How do they compare?

Codex offers a -spark model that runs on Cerebras. Not quite 17k tok/s, but _very_ fast nonetheless. Worth a look.

Maybe the 4.7 inference is cheaper than 4.6? Just guessing.

They've been rendering crawled pages using Chromium for many years now. Hidden text does not work as a ranking manipulation tactic.

How does that work?

The editing and added audio certainly add to the realism, but this is really impressive nonetheless.

Agreed. Big tech is trying to become the media industry.

Thanks, I was not aware! I thought they offered all their models via their API.

From the abstract:

> This photonic-architectured cement achieved a temperature drop of 5.4°C during midday conditions with a solar intensity of 850 watts per square meter. This supercool cement featured intrinsic high strength, armored abrasive resistance, and optical stability, even when exposed to harsh conditions, such as corrosive liquids, ultraviolet radiation, and freeze-thaw cycles. A machine learning–guided life-cycle assessment indicated its potential to achieve a net-negative carbon emission profile.

Afaik, there is currently no "GPT-5 Pro". Did you mean o3-pro or o1-pro (via API)?

Currently, GPT-5 sits at $10/1M output tokens, o3-pro at $80, and o1-pro at a whopping $600: https://platform.openai.com/docs/pricing

Of course this is not indicative of actual performance or quality per $ spent, but according to my own testing, their performance does seem to scale in line with their cost.

o3 can be re-enabled in the settings ("Show additional models") if you're a paid (Plus) user.

I think it's more appropriate to compare GPT 5 Thinking to o3. You will find that the response times are actually quite similar (at least in my experience over hundreds of identical prompts with each model).

Yup, ASP's "__VIEWSTATE" hidden form parameter comes to mind. It was base64-encoded and POSTed because it could get loooong (hundreds of KB).

Terrible for browser navigation/refresh though, because pretty much everything was a form POST. Thus no URL state sharing, either.

> merdeitocracy

Not sure if typo or intentional (likely?), but that's an amazing new word.

It's AI slop. In fact, most (if not all) of this blog's recent posts are AI slop.

I had no idea people could earn Continuing Education Units for blogging, but it's true: https://www.comptia.org/continuing-education/choose/renewing...

No idea if intentional or not, but the reason is this noindex directive:

    <meta name="robots" content="noindex, noarchive, noimageindex" />

See: https://youtu.be/f_ggYP-VO2s?t=45

It really is something. 40 FPS on a 4090, damn.