porphyra

From a hacker's perspective, yeah the feeling of accomplishment in figuring out something manually is great. But from the end user's perspective, if I get to choose between "shit ain't working" vs "it's working flawlessly but AI was involved in getting there", I'm picking the latter every single time.

too kiasu

Funnily enough now you can get the billion dollars from NVIDIA themselves.

World Labs is really cool. The 3D reconstruction from only a handful of frames is on another level.

meanwhile the Bigscreen Beyond 2:

* Display: also micro-OLED

* Resolution: 2560 x 2560 per eye

* Angular resolution: 32 PPD

* Field of view: 116°D, 108°H x 96°V

* Chipset: none lol plugs into your computer

* Tracking and sensors: needs base stations

* Weight: 107 g

* Passthrough: no

* Price: $959 (+$200 for eye tracking)

* needs base stations, a PC, and has a cable, and looks kinda like dorky swim goggles (but still pretty futuristic and cool)

meanwhile the Steam Frame:

* Display: LCD

* Resolution: 2160 x 2160 per eye

* Angular resolution: 20

* Field of view: 110 degrees

* Chipset: Snapdragon 8 gen 3

* Memory and storage: 16GB of ram with 256GB UFS plus microSD

* Tracking and sensors: inside out camera based tracking

* Weight: 185 g core headset, 440 g including strap, rear battery, facial interface, and audio

* Passthrough: idk

* Price: $1059 for 256GB, $1299 for 1TB

This is hilarious of course but Opus 5.5 just came out and largely fixed a lot of the Claudlish.

It's kinda sad that this cool project is mired in controversy. On one hand it is kinda radioactive for a former Apple engineer to work on this, even if you worked on something else entirely while at Apple. On the other hand I really, really want proper Linux support with full GPU support on modern Macs and I feel like AI is an incredibly useful tool towards that effort.

Oh well, if this doesn't pan out, hopefully someone else will vibe code for a week and get it working in the near future.

Asahi is on Fedora now (and has been for a couple of years)

I meant I can understand the fact that a legal minefield exists.

I didn't mean that I understand the actual intricacies of the legal situation.

Sorry for my confusing wording.

Very sad. I really want to get Linux on newer Apple chips working flawlessly but I can understand the legal minefield about this particular work.

This is super great. The biggest pain point of Asahi Linux is how it doesn't have GPU acceleration on M3 and newer, especially now that M6 is out!

However, Asahi Linux has a strictly no-AI policy [1]. So this great work can't be upstreamed. I expect to see a bunch of AI-assisted forks that get things working smoothly on newer hardware to dominate as most people just care about getting stuff working, while only a handful of purists stick to the non-AI version running on ancient hardware.

[1] https://asahilinux.org/llm-policy/

you could also exploit the fact that some flags may appear more frequently than others and use huffman encoding or something to encode the commonly used flags in a shorter sequence than rarely-mentioned countries, and save some bits on average

It has a list of data sources here: https://www.threebodyorbits.com/about

This whole thing does have a strong "GPT-6 Astra look" with the big text on the left style. But it is indeed very cool.

Fallout 4's endgame with the autogenerated "quests" of defending settlements get boring extremely fast.

It's even worse when you see the exact same animal asset but scaled differently with a different color hue on a completely different planet. Ruins the immersion immediately. Obviously, it has come a long, long way since the 2016 article, but still.

Also, despite the Honda being a rebadged Chevy, the average buyer may not be aware and may be drawn to the Honda brand's reputation for reliability and quality.

They can host larger models by pipelining it on multiple wafers. Each wafer stores one layer and N layers can serve an N * 44 gb model with N concurrency. The limitation would of course be inter-wafer I/O, which my comment was getting at. That's probably how they can serve bigger models like GPT 5.6 Sol [1].

[1] https://www.cerebras.ai/blog/accelerating-gpt-5-6-sol-ultraf...

Why do they only host small models rather than the 2.4T version? Is the I/O and interconnect between the wafers bad due to the limited beachfront relative to the massive size of the chip?

It seems that a rogue engineer poisoned the prompt in that instance. But the fact that they keep the system prompt open is nice. Generally I am biased towards favoring more freedom and openness rather than clamping it down in the name of safety.

The canonical view of a bicycle is facing right. Usually, people want to draw/photograph/depict the side of the bicycle with the running gear, which is on the right side of the frame for historical reasons.

Dario's "ethical" look is also kinda sus. I hate to use ad hominem, but the dude's wife literally pitched a porn film to Epstein even after he was a convicted registered sex offender [1]. Dario is also really sinophobic (it is commonly claimed in Chinese AI circles that his former employment at Baidu triggered him so much that he harbors a personal grudge against the entire race).

[1] https://www.forbes.com/sites/alisondurkee/2026/08/14/who-is-...

In my personal opinion (this will be controversial and feel free to disagree): Elon is the best.

* great contributions to many industries including spaceflight, electric cars, and self driving cars. It doesn't even matter if he is the technical mind behind these achievements or if he is just a buffoon that pretends to know the implementation details; the dude has a way of bringing together experts, having the overall vision, and managing them properly to ship amazing stuff.

* sane and reasonable takes on AI/LLM stuff. I can't really argue with "pursuit of truth" as the guiding principle. Grok talks normally without "Claudlish", has a balanced score on political bias unlike other models, has a low hallucination rate, is the best at dealing with latest news (unlike ChatGPT that refuses to believe new developments and gaslights the user), and they "never silently downgrade intelligence or fall back to other models."

In contrast, while Dario is doubtless a super smart pioneer in the AI space, his sanctimonious "We know what's good for you" attitude and extreme censorship is really offputting. The lengths to which he tries to ban or hamstring open models seems like an underhanded way to defeat competition. If he were to succeed, it would be a big setback to the thriving ecosystem of open models and hamper the development of the entire industry.

Better than even the Chinese models? That's a difficult-to-quantify, extremely rapidly moving target. Just today, Qwen 3.8 Max 0902 came out with a huge improvement over the previous Qwen 3.8 Max.

With the same logic there can never be another browser engine ever again. Which I think is actually true as everyone is converging on Blink/Chromium but it is depressing to think about.

In theory, running Apple Silicon apps on ARM64 linux machines would be really nice. But we are still far from having a performant and feature-complete implementation of doing so.

Darling only targets x86_64 stuff. Still, it's a cool project.

Kinda sad that there are very few updates though [1]

[1] https://blog.darlinghq.org/

Vibe coding is obviously not quite there yet right now. But in the future it will be. The output of agents will be just as correct and rock solid as the software from the most experienced software developers. We'll also have the additional advantage of there being existing browsers to use as an oracle to test against to ensure that nothing regresses.

I feel like with the advent of vibe coding, the bar to making your own browser is gradually lowering. People are already making all sorts of useful "houseplant programming" to serve their own pain points. Sooner or later, agentic coding will actually become really good and people will be able to just vibe code a whole browser in Rust or something that addresses the exact pain points that power users have.

theoretically equity trading is about distributing stonks most optimally so that productive companies get more and less productive ones get less, improving productivity of society as a whole

Also right now Sol 5.6 Max is super slow but if it were way faster on a chip (like Taalas' Llama 8b demo) then it would be an extreme value multiplier. But the model is so large that "baking it onto a chip" doesn't seem straightforward.

Mini split + ductless ERV [1] would probably be cost efficient and effective for this home office. No need to get a full ducted hvac setup.

[1] e.g. https://www.pioneerminisplit.com/products/pioneer-ecoasis-50...