> I have built features in 2 weeks that would take me a month just because I'd have to learn some nitty technical details that I'd never use again in my life.
In the bucket of "really great things I love about AI", that would definitely be at the top. So often in my software engineering career I'd have to spend tons of time learning and understanding some new technology, some new language, some esoteric library, some cobbled-together build harness, etc., and I always found it pretty discouraging when I knew that I'd never have reason to use that tech outside the particular codebase I was working on at that time. And far from being rare, I found that working in a fairly large company that that was a pretty frequent occurrence. E.g. I'd look at a design doc or feature request and think to myself "oh, that's pretty easy and straightforward", only to go into the codebase and see the original developer/team decided on some extremely niche transaction handling library or whatever (or worse, homegrown with no tests...), and trying to figure out that esoteric tech turned into 85% of the actual work. AI doesn't reduce that to 0, but I've found it has been a huge boon to understanding new tech and especially for getting my dev environment and build set up well, much faster than I could do manually.
Of course, AI makes it a lot easier to generate exponentially more poorly architected slop, so not sure if in a year or two from now I'll just be ever more dependent on AI explaining to me the mountains of AI slop created in the first place.
It’s too bad, really. While it’s easy to get discouraged about such things, over the course of my career all that learning of “pointless” tech has made me a much better programmer, designer, architect, and troubleshooter. The only way you build intuition about systems is learning them deeply.
Woah, so cool when a topic I was going into in depth gets to HN.
I'm a relatively new adult beginner on the violin, and one of the fascinating (and extremely difficult) things about un-fretted string instruments is the player has the freedom to shift the tuning around to fit the context. On the violin, we normally play melodies and scales using Pythagorean tuning (which is actually a misnomer as Pythagoras didn't invent it, the ancient Mesopotamians did), which is based on the circle of fifths and leads to wider whole steps and narrower half steps than equal temperment tuning. But then for double stops (i.e. chords), and especially when playing in a string quartet, just intonation, which is based on the harmonic series, is used so the notes sound concordant. This page describes all the different tuning systems a violinist may use, also including 12 TET when trying to match a piano: https://www.violinmasterclass.com/posts/152.
This video shows how challenging it can be when trying to adjust intonation when playing in a string quartet: https://youtu.be/Q7yMAAGeAS4 . Interestingly, the very beginning of that video talks about what TFA discussed that when you tune all your strings as perfect fifths your major thirds will be out of tune.
I'll also put in a plug for light note, an online music theory training tool that was mentioned on HN a decade ago: https://news.ycombinator.com/item?id=12792063 . I'm not related to the owner in any way, I just bought access a few years ago and think it was the first time I really understood Western music theory. The problem with music theory is that the notation is pretty fucked up because it includes all this historical baggage, and lots of music theory courses start with what we've got today and work backwards, while I think it's a lot easier to start with first principles about frequency ratios and go from there.
Other notes (pun intended!): The violin is great for learning music theory because you can actually see on the string how much you're subdividing it - go one third of the way, that's a perfect fifth, go halfway, that's an octave, etc. Harmonics (where you lightly touch a string) are also used all the time in violin repertoire. Finally, the article mentions Harry Patch, but you should also check out Ben Johnston, a composer who worked with Patch and was famous for using just intonation. Here is is Amazing Grace string quartet, and you can really hear the difference using just intonation: https://youtu.be/VJ8Bg9m5l50
Yes, of course there are bad actors, but this is false equivalence to equate science and the scientific method with basement randos.
Most importantly, most people don't understand scientific consensus vs. individual research papers or individual scientists. A major feature of the scientific method is that when an interesting result is published, it can be independently verified by lots of other researchers, and if they come to the same conclusion, that is excellent evidence that the result accurately describes the real world.
Scientists are people, and just like people everywhere they have biases and personal motivations. But again, the scientific method is much bigger than any individual or even group of scientists. If anything, being skeptical of unexpected results is a huge pillar of the scientific method. But skepticism alone is not enough - the next step is to look for validating research, not to say "hah, science is bullshit, let's trust this YouTube rando instead." As usual, I think Jessica Knurick does a great job explaining things: https://open.substack.com/pub/drjessicaknurick/p/trust-the-s...
True, and personally, I don't believe climate science is affected by bias to such a degree that the overall conclusion is wrong. But it absolutely does occur that a whole field can be biased, so the "independent verification by lots of other researchers" will cast unreasonable skepticism on results they dislike, while letting results they like pass with cursory examination. This is the case in e.g. social science:
The authors also submitted different test studies to different peer-review boards. The methodology was identical, and the variable was that the purported findings either went for, or against, the liberal worldview (for example, one found evidence of discrimination against minority groups, and another found evidence of "reverse discrimination" against straight white males). Despite equal methodological strengths, the studies that went against the liberal worldview were criticized and rejected, and those that went with it were not.
Man, if you are going to try to attack the credibility of the field of a hard science like climate science, try doing it with claims directly related to that field of science.
Substituting in social science as a proxy for your criticism takes the wind completely out of your sails.
"Physicists are super untrustworthy and biased, it's a cabal, I mean just look at astrology and these articles criticizing it!"
> if you are going to try to attack the credibility of the field of a hard science like climate science
But I'm not. In fact I said as much. If it'll stop you from fighting phantoms, I'll make it explicit: I'm quite certain anthropogenic climate change is real, and that climate science is broadly correct about it. Yet, not even physics is fully immune from such bias, according to Feynman: https://hsm.stackexchange.com/questions/264/timeline-of-meas...
(Though as the charts show, in physics it has a short half-life, at least for something as straightforward as the electron charge.)
What do you mean by these statements if not exactly to cast doubt on and paint the field of climate science and the consensus behind anthropogenic climate change as being the result of bias?
> affected by bias to such a degree
Implication that there is bias, and the degree to which is left up to the reader's imagination.
> the overall conclusion is wrong
What about the specifics of the conclusion, which specifics, what percentage of the assertions, 20%, 50%, 80%? This again allows the reader to fill in the blanks with their own biases which are likely far less rigorously tested than the conclusions of the field of climate science.
> it absolutely does occur that a whole field can be biased
This statement, on the topic of climate change and climate science, immediately following your above two statements, serves to further reinforce the idea that climate science is biased.
> the "independent verification by lots of other researchers" will cast unreasonable skepticism on results they dislike, while letting results they like pass with cursory examination.
The quotes around the independent verification of researchers serves to undermine their work and cast doubt on it. You then state they are unreasonable in their skepticism of results they "dislike", implying these are emotional decisions rather than empirical measurements of reality.
Of course, extremely ironic as the reader meant to consume this is of course the one actually looking for emotional reinforcement of their preconceptions, but in this framing gets to project that onto the scientists.
And yeah, "cursory exmination" of course in no way reflects the reality of the last several decades of climate science, but is added in as another unsubstantiated slight.
> case in e.g. social science
And then, the coup de grace, attempting to substitute the reputation of the famously soft and hard to replicate social sciences for climate science, in an attempt to equate the two and thus further degrade the perception of climate science.
> Implication that there is bias, and the degree to which is left up to the reader's imagination.
I would be shocked if there was zero bias - the field is staffed by humans, and has political implications. And no, I did not leave the degree of bias up to interpretation - I set an upper bound to it, that precludes global warming skepticism.
> What about the specifics of the conclusion
To date no field has been 100% correct in everything. I already told you I'm not a global warming skeptic - what do you want, for me to pretend climate science is infallible for the benefit of morons that want to twist my words?
> And then, the coup de grace, attempting to substitute the reputation of the famously soft and hard to replicate social sciences for climate science, in an attempt to equate the two and thus further degrade the perception of climate science.
On two separate occasions I explicitly wrote I believe climate science. Obviously my attempt at imparting a nuanced understanding of scientific fallibility is wasted on someone that doesn't even bother to read my posts. You want a PR statement aimed at reassuring the lowest common denominator that the scientists know what they're doing, not a discussion.
If this is how much you argue with someone who agrees with you, then, I don't know what to say. Good luck in life, man.
You must understand the net content and impact of your messaging, which is far from "Hey I'm just pointing out that humanity is fallible, apropos of nothing."
It's not -
"hey, we can argue about the best way to address climate change and the details of how it's going to play out"
it's -
"this entire field is biased" (you said "it's absolutely the case that entire fields can be biased"), the "independent verification of empirical data is actually untrustworthy and primarily motivated by personal dislike", "they make their scientific conclusions with cursory examinations", and "they're as reliable as the social sciences".
I'm sorry, but it beggars belief that you are not aware of what you're doing.
It's not the communication style of an engineer just trying to be technically correct, it's filled with subtle and not-so-subtle accusations and implications all driving in a single direction which is the discreditation of the entire scientific field.
So you think, but you're wrong. What really drives distrust in science is claims of infallibility and demands for total trust, so the only ones that remain expressing doubts are oil lobbyists and those whom they've convinced, hence your misidentification of me as one of them.
Let's take a small detour. In the show Downton Abbey, there's an episode where a character may be sick. Two doctors give their diagnoses. One says she's not sick, and that he's 100% certain. The other says he thinks he knows what illness she has, but he's not totally sure - despite knowing the patient better. Still he recommends treating her for that illness. They go with the first doctor, he turns out to have been wrong, and the patient dies.
Even the humanities and English literature graduates that wrote that show understood what you do not, that the person open about their uncertainties and faults is more trustworthy than the one claiming to be perfect. Not only is more trustworthy, but is the one that people will trust, that is why they gave that characteristic to the doctor they wanted the audience to side with - the likable permanent character, not the antagonizing guest character.
This is all a straw man, of course, as I've never said or even implied that scientists are infallible.
That said, to address your comment in some detail:
> What really drives distrust in science is claims of infallibility and demands for total trust
This is one hypothesis. Here's another -
If we look at COVID and the anti-vax movement in general, it seems apparent that a much greater driver of that distrust is grifters and kooks actively and profitably pushing emotionally satisfying misinformation. Ironically enough, they do this with far greater claimed certainty about the veracity of their conclusions than the doctors, with far less data to substantiate those conclusions.
This is pushed out to an audience which is desperately looking for explanations for the unexplainable like why their child got autism. Reasoning and nuance goes out the window and they latch onto anything that will give them a sense of certainty.
A second element of this, is that these conspiracy theories are specifically intended to stir up distrust of scientists, "educated elites", and public institutions, and to exploit the pre-existing distrust of the aforementioned that has been steadily cultivated by the right over decades.
Finally, in the case of COVID especially, we had a whole bunch of selfish motivations that were at play which resulted in people finding it expedient to believe anything that would align with their desires to do whatever they like without anyone telling them what they can and can't do (especially business owners whose financial incentives all aligned with not shutting down, and being in positions of power, they leaned into boosting this messaging).
The CDC and the medical field has for decades very clearly communicated the statistical nature of their specialty, the rate and severity of side effects, the uncertainty of prognoses and outcomes of treatments, etc etc. Quite painstakingly. There was no industrywide claim of infallibility, and yet here we are.
> Downton Abbey [...] the first doctor, he turns out to have been wrong,
Love Downton Abbey!
Climate scientists constantly communicate error bars around their measurements and the outcomes of climate change. Creating a model that predicts with any certainty how all the countless systems of the Earth will interact in such a dynamic process is insanely difficult, especially as numerous feedback cycles individually look poised to accelerate the whole process.
These many, many, many measurements and models all tend to converge around certain facts though, which leads to the probability of those being the case being as close to truth as science is capable of getting.
As "true" as 9.8 m/s^2, the heliocentric nature of our solar system, the spherical nature of our planet, the speed of light in a vacuum, plate tectonics, evolution, germ theory, etc. etc.
These truths are that the greenhouse gases in our atmosphere have been steadily rising since industrialization, that this is primarily driven by anthropogenic emissions, and that as the concentration of these gases increases, more energy from the sun will be trapped within our atmosphere resulting in a heating world.
This heating world is already, and will continue to, melt the ice caps, reducing the albedo of the poles as the darker water underneath is revealed and begins to absorb more of the sun's energy, accelerating the effect.
This heating of the world is again already, and will continue to, melt the permafrost which stores ~1.4 GT of methane, which is ~80x as potent a GHG as CO2, and whose release would represent a ~2x increase of the concentration in our atmosphere, again further accelerating the process.
This heating of the world is again already, and will continue to, lead to tropical forests like the Amazon beginning to emit more CO2 than they absorb, again further accelerating the process.
Etc. etc.
The consequences of all this are already, and will continue to, drive more and more extreme weather, with the 10 hottest years on record all happening in the last 10 years. This has negative consequences for food production, for the habitability of the equatorial regions, and will lead to mass migrations of people unless we think they will be content to just sit around and die, and thus destabilization of societies.
While we're on the topic of entertainment media another one came to my mind, it's called "Don't Look Up".
As opposed to your Downton Abbey example, the meteor turns out to be real, and it does hit the Earth, and everyone dies.
The reason humanity doesn't take this seriously and unite to address it is because there are countless predictable incentives to distrust the scientists and ignore the inconvenient truth of an incoming meteor. Financial incentives for the rich and powerful, engagement (and ultimately financial) incentives for the grifters and kooks pushing distrust of the scientists, a deep emotional desire in much of humanity to outright reject the horrors of this potential reality...
Anyway, not sure what made me think of that or how that's at all relevant here, but you should watch it if you haven't!
But besides Sam Altman, this whole episode has made me totally and completely lose all respect for Paul Graham. I used to really idolize pg, and I really used to like his essays, but over the years I've found his essays increasingly displayed a disturbing lack of introspection, like they'd always seem to say that starting a startup is the best thing anyone can do, and if you're not good at startups then you kind of suck.
But his continued support of Altman in this instance (see https://x.com/paulg/status/2027908286146875591, and the comment in that thread where he replies "yes") is just so extra disappointing and baffling. First, his big commendation for Altman is that he's doing an AMA? Give me an f'ing break. When someone is a great spin doctor I'm not going to commend them for doing more spinning. It's like he has total blinders on and is unwilling to see how sama's actions in this instance are so disgusting and duplicitous. Maybe subconsciously he knows he's responsible for really launching sama into the public consciousness, so he now just is incapable of seeing the undeniably shitty things sama has done.
Oh well, I guess it's just another tech leader from the late 90s/early 00s who has just shown me he's kind of a shitty person like a lot of us.
Yeah he has some great essays but also some that I find really dumb. Reading “Founder Mode” is when I realized he’s just as susceptible to fallacy as the rest of us.
TBH, while I may find the output style somewhat infomercial-ly, I don't really get the hatred. ChatGPT IS NOT AN ACTUAL PERSON. Like why do people care so much? Like you said, I just ignore the "persona" phrases, and just use ChatGPT (or, used to anyway, before switching to Claude because OpenAI leadership can suck it) to get information and answer my questions.
Seriously, though, just stop using ChatGPT in any case, there are very good reasons to boycott it and there are other alternatives. Not saying the alternatives are saintly, but they're not as awfully duplicitous as OpenAI.
Because people just copy/paste that shit pretending it's their own or turn their own human writing into reproduced llm text so you don't even know if they even mean what's written
It took exactly 24 hours, to the minute, from the time I received the "we're generating an export" file until I got the download link, so guessing they're either batching it or deliberately sending after 24 hours because it adds friction to the account deletion process.
I think it's pretty obvious when betting on events that are inherently just decisions by one or a few people (e.g. when will Trump launch an attack on Iran, when will a company launch a new product, will some company acquire another one, etc.) that they will attract insider trading and corruption by their very nature - all that's necessary is to have information about the decision maker. This is fundamentally different than events that are subject to forces that no single individual controls - e.g. who will win an election, where will a crypto price be in a year, movie box office results, etc.
I think betting on "single decision maker" events is just a "sucker is born every minute"-type bet.
In the past I've felt like some of the anti-Altman rhetoric on HN was overkill. It some cases it felt like piling on, and while there was definitely some shady stuff in the past, it seemed like folks were too quick to paste the "evil" banner on anything they disagreed with.
I was wrong, and I no longer think that. I now lump him in with the rest of the narcissistic sociopaths I see with so much power in the country. I'm honestly really curious what past Altman champions like paulg think of him now. I just don't see how this is the slightest bit defensible.
The "We Will Not Be Divided" pledge at https://notdivided.org/ (and discussed at https://news.ycombinator.com/item?id=47188473 ) has 96 OpenAI signatories. Time for these people to show if their signatures actually meant something or were just meaningless theater. It's not like these people would have much of trouble getting jobs given they're AI experts with resumes to back it up. Signing that pledge and then staying at OpenAI after this would just look like rank hypocrisy to me.
In the bucket of "really great things I love about AI", that would definitely be at the top. So often in my software engineering career I'd have to spend tons of time learning and understanding some new technology, some new language, some esoteric library, some cobbled-together build harness, etc., and I always found it pretty discouraging when I knew that I'd never have reason to use that tech outside the particular codebase I was working on at that time. And far from being rare, I found that working in a fairly large company that that was a pretty frequent occurrence. E.g. I'd look at a design doc or feature request and think to myself "oh, that's pretty easy and straightforward", only to go into the codebase and see the original developer/team decided on some extremely niche transaction handling library or whatever (or worse, homegrown with no tests...), and trying to figure out that esoteric tech turned into 85% of the actual work. AI doesn't reduce that to 0, but I've found it has been a huge boon to understanding new tech and especially for getting my dev environment and build set up well, much faster than I could do manually.
Of course, AI makes it a lot easier to generate exponentially more poorly architected slop, so not sure if in a year or two from now I'll just be ever more dependent on AI explaining to me the mountains of AI slop created in the first place.
reply