shaldengeki

> Is anyone really using bazel outside Google in any meaningful capacity?

Yes. For instance, Stripe uses Bazel internally for ~all of its builds. https://stripe.com/blog/fast-secure-builds-choose-two

For other users, you might peruse the Bazelcon 2025 schedule, which happened earlier this month: https://bazelcon2025.sched.com/

I appreciate your position, but the record is pretty clear; his actions clearly meet Berkeley's own definition of research misconduct.

You may be interested to read that the Why We Sleep guy committed research misconduct in the book. Guy is totally unrepentant about it.

https://statmodeling.stat.columbia.edu/2020/03/24/why-we-sle...

HN's moderation team has been pretty clear that generated comments aren't welcome here. https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

I don't see any network analysis on this page. What network analysis do you see?

I do see generic statements like "boosting each other", and I see vaguely-drawn lines in the primary diagram with no further explanation, but that hardly counts as network analysis, right?

Further in the thread, the guy notes that this isn't "new" mathematics - a better proof with tighter bounds was published in April:

https://xcancel.com/SebastienBubeck/status/19581986678373298...

Should be in the first, not seventh paragraph: this was a survey of 19 doctors, who performed ~1400 colonoscopies.

This shows a huge surge starting in 2023. I see you're counting all .AI TLDs; how much is this responsible for the surge? I think .AI TLD registrations took off starting in 2023, and one thing I wonder is if prior to 2023 we're mostly missing real AI Show HN entries, and afterwards we're mostly catching them.

I think it's important to clarify that it's just audio and video that are E2EE, not text messages themselves. (You may have meant this, but "channels" was a little ambiguous.)

Yes, that's correct. The authors themselves are being extremely careful (and, I'd argue, misleading) in their wording. The right way to interpret those words is "this is literally a model that supports our predictions".

Here is the primary author of the timelines forecast:

> In our website frontpage, I think we were pretty careful not to overclaim. We say that the forecast is our "best guess", "informed by trend extrapolations, wargames, ..." Then in the "How did we write it?" box we basically just say it was written iteratively and informed by wargames and feedback. [...] I don't think we said anywhere that it was backed up by straightforward, strongly empirically validated extrapolations.

> In our initial tweet, Daniel said it was a "deeply researched" scenario forecast. This still seems accurate to me, we spent quite a lot of time on it (both the scenario and supplements) and I still think our supplementary research is mostly state of the art, though I can see how people could take it too strongly.

https://www.lesswrong.com/posts/PAYfmG2aRbdb74mEp/a-deep-cri...

Here is one staff member at Lightcone, the folks credited with the design work on the website:

> I think the actual epistemic process that happened here is something like:

> * The AI 2027 authors had some high-level arguments that AI might be a very big deal soon

> * They wrote down a bunch of concrete scenarios that seemed like they would follow from those arguments and checked if they sounded coherent and plausible and consistent with lots of other things they thought about the world

> * As part of that checking, one thing they checked was whether these scenarios would be some kind of huge break from existing trends, which I do think is a hard thing to do, but is an important thing to pay attention to

> The right way to interpret the "timeline forecast" sections is not as "here is a simple extrapolation methodology that generated our whole worldview" but instead as a "here is some methodology that sanity-checked that our worldview is not in obvious contradiction to reasonable assumptions about economic growth"

https://www.lesswrong.com/posts/PAYfmG2aRbdb74mEp/a-deep-cri...

No, you're wrong. They wrote the story before coming up with the model!

In fact the model and technical work has basically nothing to do with the short story, aka the part that everyone read. This is pointed out in the critique, where titotal notes that a graph widely disseminated by the authors appears to be generated by a completely different and unpublished model.

> One of the AI 2027 authors joked to me in the comments on a recent article that “you may not like it but it's what peak AI forecasting performance looks like”. Well, I don’t like it, and if this truly is “peak forecasting”, then perhaps forecasting should not be taken very seriously. Maybe this is because I am a physicist, not a Rationalist. In my world, you generally want models to have strong conceptual justifications or empirical validation with existing data before you go making decisions based off their predictions: this fails at both.

This is really, really great. I hope it gets as many eyeballs on it as AI 2027 did.

The same post on the author's Substack is here: https://titotal.substack.com/p/a-deep-critique-of-ai-2027s-b...

This strikes me as disingenuous:

> I like when my software gets used by a lot of people and people send me nice emails. In this case, people weren't sending me the nicest emails, unfortunately. But they also didn't really know what I was doing. They saw DOGE, weren't a fan of certain things that they were associated with. But I think at the end of the day, like, the role of the U.S. Digital Service is to improve the UX (user experience) of being an American, which is pretty exciting. And anyone who lets me do that, I will try to work for, even if my friends and family aren't huge fans.

People weren't mad at him because he was trying to improve UX. They were mad at him because DOGE was doing a bunch of other stuff like "harassing government employees who talked about climate resilience", which he actively contributed to but conveniently omits from the interviews he's been giving recently:

https://github.com/slavingia/va/blob/main/eos/analyze_eos.py...

This is an extremely minor point, but:

> significant parts of our federal government have abandoned key precepts of outcome-driven problem-solving. (Just try navigating the IRS’ Free File tax return tool. I gave up and paid for H&R Block instead.)

It seems deeply weird to me to pick this example, of all things, where things are getting better because of outcome-driven problem-solving. (You might also want to consider that "I gave up on Free File" might score you anti-points with your audience.)

The title here buries the lede, IMO. Quoting from Epoch AI in the thread:

> We were restricted from disclosing the partnership until around the time o3 launched [...] Our contract specifically prevented us from disclosing information about the funding source and the fact that OpenAI has data access to much but not all of the dataset.

> Regarding training usage: We acknowledge that OpenAI does have access to a large fraction of FrontierMath problems and solutions, with the exception of a unseen-by-OpenAI hold-out set that enables us to independently verify model capabilities. However, we have a verbal agreement that these materials will not be used in model training.

Here, you can read "large fraction" as meaning "everything but the holdout set" - and my understanding is, they haven't disclosed performance on the holdout set.

I don't think this article, or the underlying claims, are (yet) credible. I'm disappointed that the Washington Post wrote about this at all, and that people I respect a lot have amplified it on Twitter.

The claims come from an MBA student graduating from Stanford this year, who plans on starting a developer tooling company. [1] The faculty member (a professor of psychology) who supported this work doesn't comment at all on the claims made by the student; their single quote in the article instead emphasizes how difficult it is to measure productivity.

The claims are based on, from their own description, unpublished ongoing research. There's no description of what data they gathered, what their methods were, or how they assessed accuracy anywhere. (For starters: how did they identify who a software engineer was in each of the companies in their dataset? How did they exclude data scientists, or technical writers, or other engineering-adjacent folks who in some companies need to commit infrequently? How did they identify remote workers, and handle workers transitioning between roles or working arrangements?)

Without this extremely basic information, I think it makes sense to just ignore it. It's possible that it'll eventually pan out as real! But definitely not yet. Wild that people are taking it seriously.

[1] https://poetsandquants.com/2023/02/21/from-middle-school-dro...

Here's a 14 year-old series showing that providers are the biggest chunk of the overspending in the system:

https://web.archive.org/web/20210421025041/http://theinciden...

Their commitment to open source, however, might go.

Quite recently Google quietly unshipped an effort to make their protobuf build rules more useful to OSS users of Bazel (see the rules_proto repository). This wasted a huge amount of planning and work that'd gone into the migration.

And the fact that these tools are designed first and foremost for Google use shows up everywhere. Stuff that Google fundamentally doesn't care about but is widely used (eg Ruby) is stagnant.

In this state, it's totally reasonable to reconsider whether these tools are worth building on top of. I personally still believe! But I don't blame people who are skeptical.

The Internet Archive shows the progression of events pretty clearly.

- In ~May 2023, the FT transferred it to Jake Champion: https://web.archive.org/web/20230505112634/https://polyfill....

- In mid-Oct, the site stated it was "Proudly sponsored by Fastly": https://web.archive.org/web/20231011015804/https://polyfill....

- In November 2023, this was dropped: https://web.archive.org/web/20231101040617/https://polyfill....

- In December 2023, JakeChampion made a sequence of edits removing his name from the repo and pointing to "Polyfill.io maintainers": https://github.com/polyfillpolyfill/polyfill-service/commit/... https://github.com/polyfillpolyfill/polyfill-service/commit/...

- In mid-February 2024, the polyfillpolyfill account was created on Github, and took ownership over the repo.

So I think sometime between October 2023 and February 2024, JakeChampion decided to sell the site to Funnull. I think the evidence is consistent with him having made a decision to sell the site to _somebody_ in December 2023, and the deal with Funnull closing sometime early February 2024.

As someone who attended before and after the change: there was certainly no immediate effect. The consensus at the time (at least amongst my little circle of friends) was that adopting the Common App was an obviously short-sighted metrics play.

It's definitely been superficially good for my social life, but I remain very sorry that the Uncommon App is no longer around. It played a nontrivial role in my decision to apply in the first place.

I agree with the general thrust of this, but it's worth noting that the author does _slightly_ better than is typical for LLM-based analyses: they released the dataset of book-labeled posts. You can at least estimate the false-positive rate from that, by sampling the results. (You can't estimate false-negative rate, though.)

Ideally authors attempt to do some sort of validation of the results at the LLM-labeling step and present that, but that rarely happens with these sorts of posts. I think that's pretty telling.

For whatever it's worth, the person you're replying to took the courtesy of pointing you to some research they're familiar with, and your "it's just one paper" point (or really, any of your points here) would be much more convincing if you responded in kind to that courtesy.

On topics like this you can argue from first principles endlessly and never get any closer to the truth. Better to show us some data, I think.

Sure - earlier you said "if this is useful that's a really bad sign", and now you're saying "if it's necessary it's a very bad sign", which are pretty different claims. I'm mostly interested in probing the former, so if you're not making that stronger claim then I think we're on the same page.

This is really interesting to hear; I work in a team & org that I think is pretty healthy, but I've found the yearly "brag doc" exercise useful for several years running, if only to go through and remind myself of everything I worked on. I consistently find that I've done a _lot_ more than I remember, and that's both a boost and also a healthy opportunity to reflect. The artifact is then useful over the next year as a reminder.

This is in fact the focus of Julia's post (she literally says this early on), and I think it's kind of unfortunate that folks are mostly talking about using the brag doc as advocacy in the performance review process.

It'd be useful to know what site you used to verify - but if we're talking about the same site, IMO a website that presents Dan's experience sometimes, and your experience sometimes, is actively harmful.

What's most shocking to me is how much malware there is in all of this. The fact that Google et al aren't constantly in trouble for directly forwarding unwitting users to malware distributors indicates to me just how far our standards have fallen for a "good" search engine. I feel like we'd be happier with search engines that adhered to "first, do no harm" principles.