Starlink is going to be a cellular company, except instead of maintaining cheap metal frames, they have to physically launch the antennas to space every 5 years.
You can reason it out (in a way that might make Fermi proud) pretty easily: a very large AC unit can be powered by a very small (relative) solar panel, where the solar panel is maybe 20% efficient. Thus, solar irradiation is substantially higher than AC power usage.
> Here I should note that Python is not the best choice for CPU-bound software. I want to take the opportunity to learn Zig.
For optimizing CPU-bound operations in Python, there’s some low hanging fruit with numba. I would recommend this as a 5-minute solution to you limiting your algorithm because regular Python is too slow. I regularly tell people that if their Python program is slow enough to take several minutes, you could probably learn numba before it finishes.
There's some good answers here but something important is missed. Militaries spend a lot of time having soldiers shoot rounds down the range to ensure everyone is up to snuff. The engineering corps is similar but better. You can fund target practice for your army engineers by having them go build public infrastructure. It's a win for the military (practical experience), win for locals (public infrastructure), and a win for the taxpayer (two for one deal).
Even weirder: the Kickstarter campaign says it’s 4 MSPS per channel. 100kHz bandwidth with 4 MSPS per channel just doesn’t make sense. However, they have “verified” 400kHz on their Kickstarter. Not sure who verified it, but it’s verified.
The Kickstarter does have product photos of the back in a gif but be forewarned: they don’t include any chip designations.
The specs on the physical device are quite weak too. 100kHz scope, no mention of # of bits. Power supply isn’t programmable, just +-5V at 200mA. 5V PWM and “sine and triangle” is the function generator.
It’s definitely misleading to say this replaces $1000+ worth of equipment because I’m not sure you can buy equipment with this poor of specs.
A 100kHz scope, 2W fixed dual rail power supply, and a function generator that is anything but arbitrary means at the very least projects are constrained heavily. I think part of the joy of learning is going off of the beaten path a bit, and I think this is limiting enough to prevent that.
I’ve been on the lookout for cheaper stuff to recommend to friends looking to work their way down from classical software engineering towards embedded and I think this is relatively disappointing in that regard. Second hand seems like the way to go economically but I haven’t yet looked into proper sourcing for equipment.
I would say that there’s quite substantial value perhaps not to the general public but to the general HN user from their reports. I was very excited to see them have a public scan of an LFP battery because the previous scans they did of batteries were very informative on which brands to trust and the level of inconsistency that can exist. Their analysis of BYD’s battery design is very useful to someone who is looking to understand LFP and the levels of quality control necessary to have safe batteries. The reports are a bit of a “How it’s Made” but written for the target audience of engineers and has similar utility.
Heat is exactly why this is useful. A very large amount of power draw is due to the physical size of circuits. This monolithic 3D stacking should result in smaller wires for decreased parasitic capacitance. SRAM is probably the biggest winner in some respects. The paper suggests they can make SRAM smaller, and given most of a modern CPU die is SRAM I would be shocked if that wasn’t a huge thermal win alone, either by adding more cache or by reducing the size of the core or both.
I’d be interested to see if you couldn’t throw diamonds at the heat problem though. There was some recent work done suggesting diamond could help a lot with heat, but I’m unsure if it would work here.
Malicious maybe, malware no. Not leaving your password as a sticky note on your work computer is presumably also taught in those same courses. I wouldn’t call someone typing in that password malware. If IT comes around and tries the password and then forces you to reset it it’s not even classified as malicious.
I have a hard time viewing prompt injection as malware. LLMs are unpredictable and there are many different prompts that can unintentionally cause unexpected behavior. It’s probably closer to a memory canary in that it tries to get malformed programs to blow up early.
It doesn’t read like shaming to me. It’s, in the grand scheme of HN comments, definitely on the more constructive side of the criticism. Maybe it could have been reworded, but I think the author of the post could very easily find it actionable in the future. I too had to stop reading the article at that point, so I think if the author wants more people to read, my advice for them is to just write like themselves. We’ve entered the start of a new Instagram filter age where many people feel they need to have LLMs reword their writing presumably for the same reasons as the original filter age. I share OC’s sentiment of pushing against the recent trend of implicitly shaming people for their individualistic writing styles.
It’s not about preserving a unique writing style. When I see LLM writing my brain automatically discards the content of the writing. To me, seeing LLM writing is equivalent to going to a high-end restaurant and getting served on generic paper plates. Sure, the food looks perfectly fine and there is, in theory, nothing wrong with a paper plate. Once you see that paper plate, however, you will question how nice that establishment actually is, because a lack of care for the plates undermines the quality of the food. You automatically categorize all establishments that serve on paper plates in a specific category, one that might make you concerned if you will get food poisoning that night. LLM writing is exactly the same way for me. I don’t know if this LLM-assisted piece of text is actually a Michelin three star establishment or has had several heath violations in the last year. However, I didn’t pay for it, so putting in effort to determine if it’s LLM-assisted writing from an expert or just LLM slop that isn’t from the purported author at all isn’t worth the time.
I’m much more willing to read typos and bad writing than LLM writing. If I want to read the LLM rewritten version, I can run an LLM over the original writing myself. I have not yet found true that anyone is better at prompting than anyone else in a way that suggests that I wouldn’t get substantially the same results myself. Thus, I don’t think providing the version that has passed through the telephone game is accomplishing something that couldn’t be done by readers later. I have spent the vast majority of my life reading the original writing styles of people and didn’t have an issue then. I’m not convinced a problem I had was solved when we started post-processing writing with an LLM.
I think you confuse long-term planning with long-term success. Just because something is meticulously planned doesn’t guarantee the plan wasn’t flawed fundamentally.
Additionally, wireless charging is viewed as a flop? MagSafe (wireless) is a really strong product option. Lightning vs USB-C doesn’t matter, no need to fiddle with anything in the dark, etc. MagSafe for mounting in a car is also a strong offering. Most Android phones use Qi as well.
I would like to offer some additional reassurance: I send my friends articles I see on HN that might interest them. A (in my view) very good litmus test is when someone asks where I saw it, because this demonstrates some desire for continual learning. I find that anyone that asks that question seemingly trusts an interface like HN more because of it. My suspicion is that this is probably because at a certain point you see stuff like Agner Fog's work, LWN, or a number of other minimalist websites and realize that a website that is popular despite the lack of overindulgence in UI must be popular because of the content. It doesn't hurt that the best courses in my university experience have had websites that have not changed much since the late 1990s (one did change the lime green text on turquoise background on their page after the recession to a color scheme that didn't cause headaches in students).
I do find that my peers that now read HN used to be judicial about curating a Reddit feed and mostly otherwise limited on other sources. Short-form content is addictive and as nearly as unavoidable as sugar, but many of my brighter peers work on reducing that intake. Long-form YouTube is also something I find to be a marker of someone who is seeking knowledge. Many of my peers do scroll Twitter and TikTok all day, but I find that those who are easiest to chat with are those who have already scrolled HN today and want to discuss a particular article they know I would have seen. I've had conversations that start with "Did you see that TikTok?" and conversations that start with "Did you see that article on HN?" and the latter is always more engaging.
This study I feel is misleading at least at a headline.
They found 6621 studies and narrowed it down to 51 studies to analyze. This is somewhat fine as meta-studies do need to narrow from the initial scope, but this is more aggressive than I have seen in the past.
I haven't had time to view all of the studies, but from what I can tell most of the studies are somewhat specific. That is, they are asking "If we give students ChatGPT in specific situation (specialized prompts, supervised environments, etc) then can students succeed?" I worry that the headline is quite misleading in this instance. The overall study seems to ask "if we can align the stars would ChatGPT be helpful?" This is a useful question to ask, but certainly not equivalent to "is ChatGPT helping students learn right now?"
CRWD is a member of the S&P 500. Are you truly suggesting that every holder of the S&P 500 should be put in prison? Realistically, this means that nearly every American and a very large number of international average workers just trying to save for their retirement would have to go to jail. Sounds like a good way to inflate those private prison stocks.
This isn't true. H1 antagonists, which is the group of drugs commonly referred to as antihistamines, contains two subgroups of pharmaceuticals. There are the first generation antihistamines, which are generally more popular and earn the reputation of making you drowsy, and the second generation antihistamines. The second generation antihistamines are significantly more selective for the H1 receptors you want to block versus the ones in your brain. Doxylamine is a first generation drug marketed under the brand name Unisom for insomnia, whereas a common second generation antihistamine loratadine commonly includes the phrase "non-drowsy" on the box. It still increases sedation, but at a substantially lower rate than the first generation drugs.