- NickM parentI don’t think that the production model will look like that, I believe it’s a wrap. Also suspect it’s a bit of a joke on the model being called the R2 (the colors and patterns are reminiscent of R2-D2).
- Have you really never bought a product or service for some other reason than that you saw an ad for it?
People have plenty of other ways of finding out about useful products and services. You can talk to your friends and family, or go to a store and talk to a salesperson, or look up product reviews online, or even pay for something like a Consumer Reports subscription.
- I think this is a dangerously half-true way of thinking.
Yes, there are times when hard work feels great, and it's absolutely worth seeking out this kind of work.
But any serious endeavor is going to have times that are a slog, and your ability to stick with it through the bad times will very directly dictate your ability to get back to the good times.
- SWE salaries are a massive cost. Improving productivity is one way of offsetting that cost.
In a lot of businesses you get praise and look important if you’re responsible for leading a large group of highly paid employees, more so then if you have a smaller team.
Thus the motivation is frequently to spend as much money as possible, not to improve efficiency.
If you improve efficiency then maybe you just get your team size cut and people ask hard questions about why you needed all those resources in the first place.
- The trouble is that making progress in leaps is often mutually exclusive with being productive in the short term. It’s hard to think big and plan long-term when you’re constantly overwhelmed with what’s in front of you.
Slow Productivity by Cal Newport talks about this trade-off extensively and provides interesting points of reference where real famous historical figures achieved incredible things in ways that would seem slow and lazy by modern standards.
- Truly great art, the kind that expands the field of artistry and makes people think, requires creativity; if you make something that's just a rehashing of existing art, that's not truly creative, it's boring and derivative.
This has nothing to do with whether a human or AI created the art, and I don't think it's controversial to say that AI-generated art is derivative; the models are literally trained to mimic existing artwork.
- I disagree that the current generation of AI has "solved" artistic fields any more than it's solved math or programming.
Just as an LLM may be good at spitting out code that looks plausible but fails to work, diffusion models are good at spitting out art that looks shiny but is lacking in any real creativity or artistic expression.
- I think you might be missing the point of the article: the study being cited isn't trying to establish the existence of a "language brain" or a "math brain", that's just the way the headline editorialized it to help people understand the conclusions.
The conclusion of the study was that linguistic aptitude seemed to be more correlated with programming aptitude than mathematical aptitude, which seems fairly interesting, and also fairly unconcerned with which specific physical regions in the brain might happen to be involved.
- I will never buy another Tesla again for political reasons, but regarding reliability: their new models have always had reliability problems, but then reliability has always gotten much better within a year or so.
I don't know if the Cybertruck will follow the same pattern, or if the whole company has jumped the shark, but if we're looking for non-political opinions I would not necessarily write them off on quality issues alone.
- I agree the whole thing was shady, but did they really never care about solar? I thought the solar roof was announced well after the SolarCity deal, and it seemed like they were serious about ramping that up for a while (though obviously that never really went anywhere or fulfilled any of the original promises either).
- The article is flat out wrong. The reason purple and violet look similar is not a trick of the brain, and has nothing to do with the color wheel "wrapping around"; it's a natural result of the frequency response curves of the three types of cones in our eyes. The two colors stimulate our cones in the same way, so of course they naturally look similar.
Most diagrams of our cone frequency responses are subtly wrong. Diagrams typically show three separate smooth, overlapping peaks, centered around red, green, and blue. What they leave out is that our L-cones (the "red" cones) also have a separate little sensitivity bump way off in the violet end of the spectrum. So when you see violet light, it's actually stimulating both the cones that are most sensitive to red light and the ones that are sensitive to blue light. This is pretty much the same stimulation pattern you get if you send both pure red and blue light into your eyes together, which is why purple and violet look so similar.
If you Google "cone sensitivity diagram" you'll mostly find the misleading versions of the diagrams, but you can see one that includes the extra bit of high-frequency L-cone sensitivity in this paper, for example: https://hal.science/hal-01565649/file/Vienot_ConeFundamental...
- IMO the only thing that can get through is actual personal consequences for the voter themself
Well, yes. And his approval rating has been steadily declining in tandem with the stock market declines he's caused. If/when prices suddenly skyrocket because of tariffs, you can bet his approval ratings will decline further.
- Yeah, this is a limitation, but generally if you have hard constraints like that to maintain, then yeah you probably should be using some sort of centralized transactional system to avoid e.g. booking the same hotel room to multiple people in the first place. Even with perfect conflict resolution, you don't want to tell someone their booking is confirmed and then later have to say "oh, sorry, never mind, somebody else booked that room and we just didn't check to verify that at the time."
But this isn't a problem specific to CRDTs, it's a limitation with any database that favors availability over consistency. And there are use cases that don't require these kinds of constraints where these limitations are more manageable.
- What I find hard to imagine is how the app should respond when synchronisation fails after locally committing a bunch of transactions... Manual merging may be the only safe option in many cases.
Yeah, exactly right. This is why CRDTs are popular: they give you well-defined semantics for automatic conflict resolution, and save you from having to implement all that stuff from scratch yourself.
The author writes that CRDTs "don’t generalize to arbitrary data." This is true, and sometimes it may be easier to your own custom app-specific conflict resolution logic than massaging your data to fit within preexisting CRDTs, but doing that is extremely tricky to get right.
It seems like the implied tradeoff being made by Graft is "you can just keep using the same data formats you're already using, and everything just works!" But the real tradeoff is that you're going to have to write a lot of tricky, error-prone conflict resolution logic. There's no such thing as a free lunch, unfortunately.
- Placebo is an expectancy effect. I don't know all the details of OP's story, but there are all kinds of plausible reasons I can imagine that someone might have different expectations for one drug over another.
It might not even have anything to do with the drug itself: mental health issues tend to wax and wane on their own over time, so if someone happens to feel better right after starting a new medication, it's easy to think "oh hey this one must be working" and then that can trigger the placebo effect and turn into a positive feedback cycle.
- Keep in mind that studies find a strong effect for placebos: the numbers are not saying "these pills do nothing" they're saying "these pills seem to do a lot, but placebos do almost as much".
Obviously the effect feels extremely real to you, but we wouldn't see a strong placebo effect in the numbers if people on placebos didn't genuinely feel much better.
I get that it feels like the second drug worked much better, but expectancy effects and internal narratives are extremely strong, and they're impossible to untangle at the level of an individual.
- Exactly this.
Many double blind studies are completely broken due to side effects triggering a stronger placebo response, and this is an especially huge problem for drugs like SSRIs where a placebo gets you about 80% of the benefit of the actual drug.
Similar to the study you linked, there was a more recent study where they found that for the SSRI escitalopram (aka Lexapro), the benefits disappear when you lie and tell people that they're receiving an active placebo that mimics the side effects of an SSRI. That is, if people don't actually think they're taking an SSRI, they don't get any benefit.
https://app.dimensions.ai/details/publication/pub.1142338190
- I used to buy into this kind of stuff, but I've become more and more skeptical of the idea that you would still be yourself if your brain could be preserved/emulated/transplanted/whatever.
Our nervous system extends into our bodies. We feel emotions in our bodies. People with certain kinds of brain damage that prevents them from feeling emotions normally also experience trouble making rational decisions.
More recent research has been hinting that we may even hold certain types of memories outside our brains.
Humans have always been drawn to neat, tidy ideas, especially ones that draw clean boundaries: it's an appealing idea that our consciousness lives solely in our brains, and that our brains could function independently of our bodies, but it seems unlikely that it's really that simple.
- There were customers happily using strong consistency in production, but somehow the idea that it wasn’t “finished” kept getting repeated over and over by management. I was well on my way to solving the biggest rough edge (tombstone reaping in SC buckets) but then I got pulled off to work on the infamous “data platform” and never got to finish that work :-(
- I just got new garage doors installed with new openers, and I saw that the openers support myQ, and that my car does too, and I thought "sweet, I'll be able to open and close the garage from my car without any extra hardware".
Turns out myQ charges a subscription fee for the privilege of using this feature.
I could install an opener like this one, but it still wouldn't solve the in-car integration side of things. Anyone know of any clever workarounds for this? Seems like maybe I could MITM the myQ service since my car and garage would both be on the same WiFi when I'm home, but I don't know if there are OSS replacements for the myQ server software.
- Even when real wages keep up with real prices, people still hate inflation, because they attribute their rising wages to their own successes more than macroeconomic changes. To most people it feels like "I'm working hard and getting big raises for it, only to be stymied by rising prices" rather than "this is all happening due to forces outside my control".
- Unfortunately it's hardly possible to do proper case control studies with psilocybin, since the psychedelic effects cause unblinding.
This is a problem with SSRIs as well; studies have found most people can correctly guess if they’re in the study group or control group.
A true double blind study would require a drug that has similar side effects via an unrelated mechanism, but nobody has attempted that as far as I’m aware.
There was this interesting one though where they deceived patients into believing they were taking an “active placebo” instead of an SSRI, and found that the benefits went away, which is very interesting and raises a lot of questions about the true efficacy of these medications: https://www.nature.com/articles/s41398-021-01682-3.pdf
- I think a major criticism of how this study is designed is that it is not truly "blinded". I doubt many of the participants were unsure if they received psilocybin vs escitalopram.
This is already a problem for most psychiatric drug trials; studies have found that most people can tell if they’re on an SSRI or in a control group, since the side effects are pretty noticeable.
There is evidence that this skews study results, to the point that some scientists believe that SSRIs may only be “effective” because the obvious side effects make them a good active placebo.
- I'm not claiming they don't.
You were clearly implying they don’t when you downplayed all SSRI downsides as “a handful of annoying side effects”.
Weight gain kills. Suicidal ideation also kills many people when they start SSRIs. People have killed themselves because of permanent sexual dysfunction caused by SSRIs that never went away after stopping the drugs (and yes, this is a known possible adverse effect listed in the official prescribing information).
I don’t think the choice is as “obvious” as you’re making it out to be.
- There’s no technical reason different social media networks couldn’t be made to interoperate. The barriers that keep other platforms from building off of a successful network are largely legal, not natural.
I’m also equating these kinds of network effects with other artificial barriers to competition, like price fixing or mandatory non-compete agreements.
- This shows the true value of the platform to end users is near zero.
Is that really true? I think platforms have strong network effects that make migrations to alternatives (like Bluesky) very difficult.
The network effect has negative value to the user; if the platform went away, everyone would migrate to a better platform, and everyone would be better off. The network effect is an artificial barrier to competition that only benefits the owner of said network, and it only works because collective action is harder than collective inaction.