5,626 karma · joined October 6, 2015
There is one near me, in a relatively upscale area, that has repeatedly been taken down and thrown into a culvert. There are currently two old flock cameras still on the pole down there and seemingly unreachable. But it is always quickly replaced.
I think it might be worth stepping back and thinking about what the core elements of the product is. Not just with X, but with with all social media.
Personally I've migrated from thinking 'damn this is broken' to 'oh thats a feature not a bug'. I'm embarassed how long this really took me across a lot of platforms, but when I stopped thinking about my needs as driving product design decisions it all was less frustrating (albeit more disturbing).
I nearly cut contact with my mother over he wanting to use 23andme for genealogy purposes. The threat was ugly but as far as I know got her to not do it.
There are a lot of people who have now made major choices for others personal data who had little meaningful informed consent. Now it’s up for the highest bidder. I have some small hope the EFF will say something on behalf of consumers and a court will listen. I it pure hopium but it’s all I’ve got.
I have this imagine of Claude suddenly randomly asking if I want to go buy a coke.
1.5B transistors in a current intel core chip
300 die per wafer
150 wafers an hour
That means each litho tool prints 6.75 x 10^13 “transistors” per hour. In more useful units, that’s 18.75B transistors per second.
Drawing them one line at a time is technically feasible but…I’ll bet you are talking single digit DIE per hour if that.
And that is for one layer of lithography. I’ve seen estimates from 5-20 layers of lithography using EUV tools at the 3-7nm nodes. So the time scale is even more warped.
My university has an alternative entrepreneurial focused “senior design” course open to engineers and cs students. Because of scale the CS students are now separated in a different section. All I can say is the level of toxicity that I’ve witnessed from cs students to engineering students has made this a much nicer course to work with teams.
Engineering students can be ego driven now it alls sure, but the level of toxicity behavior and assumptions of superiority I’ve seen from CS students makes me glad I don’t have to deal with them anymore.
The primary fault I see, and it’s buried in the abstractions comments from others is the assumption that their knowledge is all encompassing because it is so abstracted. The inability to attach it to contextual realities is taken as virtue rather than a risk factor.
I always go back to the fast company article on the coding team behind the space shuttle and their culture. The difference is telling in terms of qyality and risk tradeoff. Realistically the number of engineers who will design something that could kill someone and software developers who will exist at opposite ends of the percentage scale. That reality manifests in professional culture. We’ve seen it in so many places.
This threat is insightful but not in the way commenters likely think.
SEC deals with oversight and regulation actions/cases against obstinant defendants all the time. That is hard because those folks know the law.
Trying to deal with a teenager or influencer who doesn't even have a concept of the securities law is not something they do regularly. It doesn't fit with their system of function, and would likely become an absolute doom spiral of wasted time and distracted effort. Imagine trying to do a deposition where every other response is 'skibidi'. Even if it worked what are they going to collect or recover? Other meme coins?
In short: Setting aside the fundamental logic of this position, the SEC is built to battle with Bobby Axelrod, not Huawk Tua girl and those are fundamentally different skillsets.
It also goes to how a lot of people misunderstand the replication crisis. 'Hard science' really should replicate - we should be able filter out sources fo error and variance because the phenomena (generally) isn't affected by our attempts to measure it. Making social science replicate often requires so much control that it is deabstracted from reality, meaning the effort at replication reduces the value and usefulness of the knowledge. Generalizable claims are hard because the sources of variance are so much larger adn more complex. Speaking as someone who went through a transition from engineering to social sciences, it is the concept that made it hard. I started my time in social sciences with a cool idea of a whole carrer based on just doing replication studies, because science. That was...useful and stupid at the same time.
The larger phones barely fit in pockets, fit poorly in running packs/biking gear and are just generally inconvienient for being active with them.
That entire second group in this article, is talking about a set of things that relate to one of two NSF review criteria. One is intellectual merit, which is pretty much what you think. The other, is what this is all related to is called or impacts. The broader impact review criteria were developed in the mid 90s on direction from Congress to ensure that science funded by the government, benefits society and the public in a set of particular ways. As with the science, PI’s get to choose what types of broader impact activities they do. As with the science itself, you and I might find some of them an interesting or even to testable, but if they meet the review effectively, the funding is not because of what they do because they met the criteria.
There are valid arguments for and against the entire world abroad impacts in NSF research. This isn’t a conversation that is part of that arguments for against. Instead, this is a child who knows nothing about it, screaming and demanding attention over people who do.
These grants were funded through normal nsf scientific processes by experts. The report refers to that funding as “diverted” because they don’t like the content, which is just manifestly dishonest. The are implying that the funding was stolen or decisions made maliciously.
That report does not have truth or valid data as an intent. His commentary at the end shows where this whole argument, and the larger rationalist movement, fails when it encounters the real world. To him the false positives are errors that affect the data and larger claims. To the people creating this false positives are beneficial because they dishonestly multiply the claims they are making. The tell is that it is so easy to do this analysis and no one bothered.
Both this response and the original report also fundamentally misunderstand how grant reviews and broader impacts work.
> It reflects poorly on the Biden administration that you could only get a grant to cure cancer if you suggested you might teach an underrepresented minority child about it.
Implying that this is specific to the Biden administration is taken as true by this re analysis. It has been a review criteria since the 90s. An honest attempt at this work (I’m not even sure that exists) begins by looking across administrations.
But honestly it’s only been in the last few years review panels began to care about broader impacts besides “yup it’s there”. Pi’s pick there own broader impacts work. Most of this is all congressional mandated.
Stop taking lying about things as counter argument because it’s not. This is why the rationalist movement will always fail in the face of reality.
1) because it can mutate and develop new and different symptoms (c.f., yearly flu)
2) because the lack of testing/symptoms means its more difficult to identify and track (c.f., early covid)
3) because people can be infectious without knowing and pass it along...see 1
4) because no symptoms now doesn't mean it does not cause harm (c.f., HPV and cervical cancer)
The immediate post graduation employment of all PhDs is very high 74% in 2023 up from 68% in 2018 and the highest since 1993. Interestingly difference between stem and non stem is effectively moot.
I continue to be shocked in this thread at the factual statements being thrown around that there is easy data to push back on from the nationally managed survey of earned doctorates.
35,566 / 57862 = 61.5% (overall)
26,622 / 45,533 = 58.5% (stem PhDs)
Survey of earned doctorates, national center for educational statistics, 2023 data…very useful, as are many of the data products the federal government collects, for however long this is up
This sounds more like an insight into how things are working at open ai than anything else. And I’m not sure if deep seek and others are going to follow his nice rules.
More generally, transistors are a technical phenomenon…they are either smaller (and work) or don’t. The thing I really don’t feel enough folks appreciate about AGI is it’s a social phenomenon - not in the making of it but in the pragmatic reality of it.
To a sufficient number of folks the current version is AGI, I see students every day trust it more than themselves. To bosses it also might be, if it’s more intelligent than you average employee than that’s sufficiently general intelligence to replace them. So far, I’ve really tried but don’t beyond generating the most basic outline I have yet to find a model that helps with my work, so it’s not intelligent for me.
I’m aware of the benchmarks but they don’t matter outside of places like HN. intelligence is and, likely, will always be social before it is technical and that makes these laws..not useful?
Your concern about propaganda rings hollow given that you are trusting a government source talking about a government source.
https://www.dw.com/en/fact-check-trumps-team-targets-usaid-w...
This simply is not the case, I know it is something that you and many others believe is that case but you are being lied to by actual racists. I say that as a white man working in STEM academia. Academia had a long history and tradition of NOT doing meritocracy, but of claiming meritocracy and using bad markers of meritocracy to prove it. The 'DEI work' that people are so concerned about is about trying to make merit based decision making actually merit based. You think that is based on giving some preference, but it isn't - its based on acknowledging and working to eliminate actual prejudice. Its about hiring the best people instead of the person who's advisor is friends with our search chair.
I'll give a concrete example: I ran a hiring search for three faculty members. We did a blind search. The hiring committee did not know the gender, race, ethnicity, or even institutional affiliation of any candidate. The candidates were ranked, the top invited for phone interviews, and then ranked again during the interviews with everyone blind to the first set of rankings. We repeated this for a smalled group of in person interviews The order of the rankings at all three phases matched. More relevant, we interviewed and hired the most diverse crop of faculty we have ever hired. Simply because of the appearance who we hired, two different candidates who were did not received interviews emailed me and my department chair to decry that we had used 'DEI' in our BLIND hiring process. One threatened a lawsuit. We blinded it to race, to gender, to all markers of 'diversity', but the gender and race of who we hired was all the proof that person needed they were less qualified than him.
In other cases, this 'stuff' protects against asshole colleagues and bad science. (1) Diversity statements help us avoid getting sued by students and employees. The statements that wax philosophic about inclusion, that quote MLK, the ones people use to label this as some ideological test, SUCK to read and get applications ignored because they lack serious thought about being a colleague. The good ones, which get noticed, are about how people work effectively with other people, and how they make an effort to understand people as part of working with them. The context, whether its about being Green, Left-handed, or Neurodivergent tell us whether this person has thought about being a mentor to people unlike them, has a capacity to empathize with a student, or is going to be a self-righteous asshole that is going to make us hate faculty meetings even more. They help us know if their grad students are going to be in tears in the chairs office or the parents of an undergrad are calling the dean. (2) They actually tell us a lot about doing good science and getting grants. Theres a long history in medicine of fucking up because of who is in our participant pool. NIH now makes you articulate about how you will not do bad science through lazy recruitment[2]. We've asked questions about this requirement of candidates during interviews. The answers are fun and telling - using coded language to say you won't recruit Black people because they are 'less reliable' is just evidence you don't get it, not that you are some purist doing important work.
That is the DEI you are being propagandized to be against - what it actually is not what you are told it is. It is not hyperbole and you are tired by design - because you are a victim of propaganda. The nonsense narrative that is being pushed is, without concern for the truth, entirely grounded in the assertion that certain groups are unqualified to do intellectual work (c.f.[3]). It is (by design) meant to establish that the mere appearance of a Black Women or a gay person on a faculty is only because they are unqualified. It is meant to exclude people who have always been excluded. It is not about pushing back on (nonexistant) out of control efforts to include them. What is changing is efforts to counteract the actual, long established, clearly evidenced, bias in favor of certain groups of candidates. That is not some ideological project to eliminate people like me because I'm not a minority, that is the thing you want.
[1] https://mbb.yale.edu/news/mbb-radically-changing-how-we-sear... [2] https://pmc.ncbi.nlm.nih.gov/articles/PMC6053906/ [3] https://www.rollingstone.com/politics/politics-news/darren-b...
I feel more and more like a large portion of the American public exists at the weaponized intersection of the subbing Kruger effect and Chesterton fence. They hear so many vague platitudes about waste that it’s just taken as dicta without evidence. That somehow provides a global mandate to break anything.
If we survive this I hope that the government workforce starts to get more respect of the hard work they do on complicated problems to make fair processes and that they stop getting just blanket accused of incompetence for ideological gain.
Lets be clear, that is not the goal - that is what they say the goal is and reality shows it is not. The goal is grift and theft adn destruction. Properly naming things is going to continue to matter more and more. Because no matter your bias or perspective, repeating propaganda is an act of propaganda.
A marketing team isn’t a generality that makes a company known, it often focuses on communicating what products different types of customers need from your lineup.
If I sell three medications:
Steve
56285
Priximetrin
And only tell you they are all pain killers but for different types and levels of pain I’m going to leave revenue on the floor. That is no matter how valuable my business is or how well it’s known.