tao_oat

Sure, but I fail to see where the theft comes in.

This is all consistent with the beliefs of many AI (and EA) types who are not associated with the company.

> effectively stole

Could you say more about this? I'm not very familiar with the WhatsApp acquisition, or necessarily happy about it, but at first glance I don't see what part would come anywhere close to theft.

Does anyone actually need IE11 support in 2026? It was released 13 years ago.

Cool to see the BBC writing about this. If you're interested in this I suggest getting involved with PauseAI: pauseai.uk

I think the problem is primarily that a superintelligence would be fundamentally inscrutable to us, i.e. we don't know how it would think or what its goals would be. It might decide that humans are a minor inconvenience to achieving its goals and thus worth removing. Or that burning all carbon lifeforms could power its GPUs for a week.

Even if it wouldn't want to do this at first, the fact that it'd have the capability to seems bad.

I believe this is how many CSAM classifiers work. There isn't a simple legal avenue to host CSAM for classifier training, for good reason.

They're fine for a CDN but overall their offering surprisingly immature IME. The devx for their hosted scripts/database just isn't there, and e.g. they only allow one global API key with full permissions, etc.

This is a surprising opinion to me! Subjectively, I very strongly feel that models improved significantly over the last two years, particularly around the time of Opus 4.6 and now Sol 5.6 and such. It's made a huge difference to how much autonomy I can give them.

If you trust benchmarks (which is debateable) then the rate of improvement is speeding up -- definitely not diminishing returns.

I think it's one of our better attempts to understand a very large complex system! I think it's inherently very hard to do. I do think you can critique a lot of things about the field and ideology, and so on, but I don't think that's an inherent problem of economics.

> non-obvious

I'd maybe question this a bit. I suspect that many things we consider obvious are only so because of the field of economics -- but that's admittedly only an initial hunch.

Do you think there's any value in trying to understand how human societies work at scale?

This is my experience too. Many people in my offline, non-techy life seem to get almost upset when I argue that models are good at many things (and getting better fast).

And I kind of get it! If your primary exposure to AI is Google search summaries (from their cheapest models) and shitty AI graphic design, it makes sense to think it's vastly overhyped.

You're skipping his recent blog posts; I think that's where the bulk of people's complaints come from. See for example https://tekin.co.uk/2025/09/the-ruby-community-has-a-dhh-pro...

dhh is very strongly associated with identity politics, though? (as in, "western"/white identity)

I've been following Turbopuffer for a while; IIRC the founder talked about scaling MySQL at Shopify. So might just be operational familiarity.

For some reason we seem to have more patience for bad writing when it's clearly from an LLM. This article starts

> US 12,670,045 B1 was filed on 4 March 2026 and granted on 30 June. The claims are now readable and they are narrow: a stateless sandbox that pauses a code block, ships one tool call to the client, and resumes by replaying from the top. The prior art that bears on that is durable execution, not CodeAct.

This is really hard to parse! What's CodeAct? This article doesn't seem worth anyone's time. (it clearly wasn't worth the author's time to write it).

Electricity (including heating) is actually the single biggest source of greenhouse gas emissions, but transport is indeed number 2: https://ourworldindata.org/emissions-by-sector

Followed by agriculture & manufacturing.

Seems incredibly unlikely, doesn't it? (I hope for the same thing but it feels almost like wishful thinking!)

Yeah, I think it's largely a UI issue! I really want a UI that lets me review distinct changes that are dependent on each other, and handle comments/changes neatly within that structure. I've found GitHub quite lacking there historically.

Also, if you're the mchristen I know, then I think we've had this discussion before haha! Hope you're well :)

I think we're saying the same thing, that's how I use stacked PRs.

Here's a common flow where I find stacked PRs are useful:

- I want to build feature X

- Ah, but it would work better if I refactored the module first

- I refactor then build feature X

- There's then some additional (and optional) cleanup work

As a reviewer I wouldn't want to see all this in a single PR, and the changes depend on each other so I can't open multiple independent PRs. Manual rebasing is fine but navigating the GitHub UI is then annoying, I have to mentally keep track of where I am in the stack.

Really excited to try this. After using Graphite it's been very hard going back to stack-less GitHub. Hopefully this can make the stacked PR workflow more common and give people an easier alternative to mammoth PRs.

> It uses the definition of "legally significant" from the GNU Project maintainer guidelines, which holds that the threshold is ""around 15 lines of code and/or text"" to qualify as significant for copyright purposes.

From the page itself, linking to https://www.gnu.org/prep/maintain/maintain.html#Legally-Sign....

I think this app is a great example of AI-driven app creation!

I generally rely on LMArena for this: https://arena.ai/leaderboard/code/webdev/pareto

But it does take some days after model release before they collect enough data.

Interesting; I was looking for a replacement for Strong for iOS just yesterday. Strong was a great app but the dev hasn't been keeping up-to-date with iOS features like live activities and such.

I found that there are loads of apps like this out there, with varying degrees of polish / paywalls. Don't get me wrong, this one doesn't look bad! But what I really want is a demonstrated long-term commitment to the app.

If I can give one piece of feedback, the LLM copy on this website doesn't inspire confidence. Like it mentions "Identical layout: kg always on the left, reps on the right." Is that really a feature? Or "Confetti, only when earned. Never as filler." -- idunno, that doesn't seem like it deserves a spot on the landing page to me.

I tried to list all the traits I could think of:

- Orange/beige-ish colors - Rounded corners - Cards with a thin border - A thicker colored border on the left of cards - Serif font for headings - Monospace fonts for small text - Headings that often have an unnecessary subheading / pre-heading - Little badges, often with a "status indicator" dot on the left - Obviously LLM-generated text / language

I'm sure I'm missing many but the above are dead giveaways!

> moving code repos into a decentralised hosting environment

It's unclear to me what Entire means by decentralized. Based on their most recent blog post (https://entire.io/blog/an-entirely-new-git-hosting-network) it seems like they just mean globally distributed, but all controlled by them.

In contrast, Tangled at least offers something where you can own your own data!

The price seems sensible to me, these AAA titles are expensive to create.

Not that it'll affect me personally; I'm not buying this game after Rockstar's [union busting][1]!

[1]: https://www.bbc.co.uk/news/articles/c9v10rr1meeo

I'm using Bunny DNS and it's been mostly unremarkable (which is a very good thing for a DNS provider)!

The only annoyance is that their domain import auto-detects existing records, but it seems to miss a lot of them so you end up manually copying a lot of things over anyway.