Never trust Google AI summaries, it hallucinates like crazy. I've recently googled my business name a few times on different machines and each one had at least one factual error, and what's worse - the errors were different for each run. Another time I wanted to know how to do something in Notion, and Google hallucinated a feature that does not exist, listing exact steps on how to enable it.
Try it - google your name, name of your company or "how to do X in software Y" and you'll see for yourself.
The diagrams were generated with Nano Banana Pro (most probably, or alternatively with ChatGPT Image 2), if you look closely in high contrast areas you'll see artifacts in the background that give it away.
I personally don't mind AI generated content when it's properly reviewed, but unfortunately more often than not the author just glances at the result and decides it's good enough.
I'm not knowledgable enough to determine whether this diagram is 100% accurate, but some things look off - the arrows in the bottom left seem superficial, some arrows are connected in weird ways, the mini diagram in AttentionLayer block doesn't look right (it has two Softmax icons and one MatMul icon, while the "before" diagram is the opposite).
Fun idea! I was surprised that it loaded on mobile and UI was kinda reasonable but I could tap any buttons to move or do anything. I’ll look into sending a PR for fixing that.
This kind of behavior should be met with swift and strong reaction from both the school authorities and the law enforcement. Kids responsible should be put on trial and face serious consequences including jail time for most egregious cases. Let the news cover it nationwide. "This is what will happen if you even think about creating and circulating nude pics of anyone" should be the message sent. This is the only way to prevent this behavior - not more censorship, not technical solutions
> What’s the point of adding features that are inscrutable?
You are assuming that the additional speed comes at a cost of codebase comprehension. For me it's not the case - I never push generated code I don't fully understand. It does take time, sure, but it still takes me much less time to write a spec, execute with AI and then review than write the thing myself.
Sibling comment didn't elaborate, but I think they might be onto something.
It happened to me personally - LLMs and agentic coding tools enabled me to pick up old side projects and actually finish them. Some of these projects were in the drawer for years, and when Sonnet 4 released I gave them another try and got up to speed really quickly. I suspect this happened to many developers.
This was the case for me a year ago. Now Claude or Codex are routinely delivering finished & tested complete features in my projects. I move much, much faster than before and I don’t have an elaborate setup - just a single CLAUDE.md file with some basic information about the project and that’s it.
Even if your tests take a long time, you can always (if hardware permits) run multiple tests in parallel. This would enable you to explore many approaches at the same time.
Upon deeper inspection? Someone checked the Cargo file and proclaimed it was just Servo and QuickJS glued together without actually bothering to look if these dependencies are even being used.
In reality while project does indeed have Servo in its dependencies it only uses it for HTML tokenization, CSS selector matching and some low level structures. Javascript parsing and execution, DOM implementation & Layout engine was written from scratch with only one exception - Flexbox and Grid layouts are implemented using Taffy - a Rust layout library.
So while “from scratch” is debatable it is still immensely impressive to be that AI was able to produce something that even just “kinda works” at this scale.
It’s not just a wrapper for Servo, the linked poster just checked the dependencies in the Cargo file and proclaimed that without checking anything further.
In reality this project does indeed implement a functioning custom JS Engine, Layout engine, painting etc. It does borrow the CSS selectors package from Servo but that’s about it.
Nothing I found says anything about Zig folks being inherently against AI. It just looks like they don’t want to deal with “AI Slop” in contributions to their project, which is very understandable.
These are ~2 years behind state of the art from the looks of it. Still cool that they're releasing anything that's open for researchers to play with, but it's nothing groundbreaking.
Is $14 dollars for ad-free, unlimited access to literally billions of videos really a steep price? Personally if I were to get rid of all but one of my media subscriptions I would stick with this one, since it's got everything - entertainment, education, inspiration, you name it.
It's only a matter of time this will be fixed, also there probably already are custom LoRAs that can remove jpeg artifacts. So it's not a matter of if, only when.
Words are cheap, but "We are sorry." is a surprisingly rare thing for a company to say (they will usually sugarcoat it, shift blame, add qualifiers, use weasel words, etc.), so it's refreshing to hear that.
