I'm the same way, I hate using Google search for searching because it's basically useless, and their other ecosystem offerings generally get enshittified over time so it's not worth paying for or relying on.
But if they're letting me using AI for free without logging in and I just need a dumb AI slop answer, then I'm more than happy to burn their tokens instead of my own. Any serious work goes to a different LLM provider. The switching cost for moving to a different LLM provider in the future is practically zero.
I was thinking maybe have those chemicals sitting in a glass or temperature sensitive container inside the tank. So when there's too much pressure or heat, the container containing the neutralizing chemical is broken like a fuse and the chemical is automatically released.
Well then... make a matrix of such fuse-containers? (say every 20cm or whatever)
I guess manufacturing such a matrix would be pretty expensive though, you'd need to carefully automate its production I think. It would also definitely interfere with flow of fluid in the tank.
I was thinking multiple long skinny tubes with etchings that make them more likely to split lengthwise. Maybe with a spring loaded/powered agitator so when the tube breaks there's some mechanical flinging/mixing of the inner chemical.
But I'm not a chemical processes engineer, so I don't know how much mixing is needed. But the existing emergency plan was to inject the the chemical through a single valve, so it seems like the dispersion and mixing requirements in this case seems to be low.
I am imagining some sort of regulations that enforce the creation and installation of appropriate safety systems upon all similar sites in the affected jurisdiction.
And a standard to prevent similar danger elsewhere in the world.
Or is that too fantastic? Perhaps I should stick with the distopian fiction I so love. Why haven't flying cars taken off yet?
I always wondered why people don't also ask the AI to generate code comments/documentation, summaries of those documentation, overview of the system, and re-review them all for correctness for the changes they asked the AI to do.
What I've noticed reviewing all my colleagues' AI generated code PRs is: it really is just code, and the rare comment here and there is still added by the human.
We're already trying to light tokens on fire as fast as possible to stay on acceptable required use leaderboards, why not light some more for system understanding and housekeeping.
The docs/summaries part I can get behind if reviewed and improved by a human, but at least when working in pre-existing codebases, I tend to steer models away from writing comments, because I find that almost all comments they write are "not even wrong".
That is: they either reiterate what the code does, or would if the code were slightly clearer, or they tell half truths that are more confusing than helpful. Mostly they fail to emphasise the salient things, like the why over the what, that are not obvious from the code.
yes it makes sense that it's AI all the way down, but also the "why not just…" answer is because it's exhausting.
Being reduced to an inconsequential middle manager is more exhausting than being reduced to a code monkey is the hot insight i've been hanging my hat on.
To be clear: This is really horrible in an IC for the paycheck role. I quit my job on principle because of code/token maxing. Very few are in the place to do this. I've been enjoying AI as an independent, but i still mean to fight the good fight for every line engineer.
> I always wondered why people don't also ask the AI to generate code comments/documentation, summaries of those documentation, overview of the system, and re-review them all for correctness for the changes they asked the AI to do.
I now on all of my projects have an ai journal that stands as a ledger for every change the ai has made, and why it was made. I don’t read it that hard personally because I spend so much time planning with my agent before letting it code. However I have found it very useful in sharing code between people, or having Claude look through the journal to gain context when modifying or adding a feature.
I see people asking LLMs to omit comments, but comments like PDL (Program Design Language) could be helpful for interactive development: ask the LLM to write pseudocode comments, review them, and then ask the LLM to expand them to actual code. People say comments should explain why, not how. That seems useful for both human reviewers and LLMs.
What harness? In my experience, Claude shoots out tons of very low-quality comments. It's always too hyper-focused on the exact specifics of the bug as presented to it, with no higher-level generalization of the concepts involved. In a big codebase, this means the comments are meaningless without a human rewrite, but it definitely writes them.
I’ve have amazing, amazing results through composing an “ADR” skill hook. It knows contextually what they are for; how to write them; when to include them as steering.
For major new features, I do ask it to do those things. Why is it that you think people don't do that?
Even for most small changes I will ask it to do a simple "production review" then I use my experience and judgment to decide on which items need to actually be addressed or not
> When exposed to a trigger -- such as a small amount of heat or a catalyst -- the molecule snaps back into its original form, releasing the stored energy as heat.
From the paper abstract, the catalyst is HCl. I don't have access to the full paper, so I don't know how they separate the HCl from the MOST to neutralize it to be rechargeable again.
From what I've seen of the paper, it seems like the catalyst is needed for a full energy release reaction. Regular batteries also have rapid energy release with unintended contamination of a chemical (water), and we still generally have no problems during regular usage.
The reaction triggered by heat doesn't release all the stored energy, which would be the bigger concern for unintended runaway reactions.
As a sibling comment says, in the paper the catalyst is acid and I can't find the release using heat in the paper. Is that an hallucination in the press article?
(Note: At high enough temperature the thermal energy will be high enough to go over the energy wall from the high energy isometric to the normal one. So it's plausible. But with high enough temperatures a lot of nasty things can happen, like total decomposition or just burning. So it may not be a good strategy to make a rechargable baterry.)
Just make the punishment the seizure and full release of the game assets (all source code, version control history, tooling, and release of copyright/trademarks).
It's always going to be a wild goose chase trying to take money when there isn't any (actually or by design), just take the product and let the public update it as a last resort.
Costco is for bulk staples and commodities for me. Products that I really don't need the best of the best for, good enough is good enough and as long as I can use it all before it goes bad, I'd rather not waste more thought than needed for it. Milk, eggs, flour, flowers, microfiber towels, batteries, salt and pepper.
Then for all the niche stuff that I do truly care about, there's the specialty stores or really the farmer's market. That's where I'll indulge for the first press seasonal olive oils, all sorts of pluot/apriplums/plumpicots combinations, short shelf life wild berries, blueberry/orange/mint blossum honey and whatnot.
> Costco is for bulk staples and commodities for me
> Milk, eggs, flour, flowers, microfiber towels, batteries, salt and pepper
If you can walk out of Costco month after month with just those essentials, and never pick up any of the nice-looking and reasonably-priced goodies there, I think it's safe to say that you have the level of discipline required to be financially successful and not have to care about whether you're getting the best price per unit of whatever it is you're buying. :-)
Kernel level anti-cheat also doesn't introduce a giant performance penalty like Denuvo-style DRM. People just want to play their games without it still stuttering on top of the line hardware.
Anticheats will still have obfuscated code for obvious reasons (they don’t want to be reversed). Not sure they don’t induce some performance drop too - though maybe smaller compared to bad Denuvo implementation.
One of the first things mentioned on that page is:
> To protect our intellectual property, certain features – such as fan impeller geometries – have been slightly modified while remaining visually very close to the actual product.
So you do have to 3d scan them yourself if you're trying to print a copy.
It's funny because I replaced my NF-A14 and NF-F12 because they had hums at certain rpms when used on radiators, and neither the Arctics before them, nor the BeQuiets that replaced them, had that issue.
I used to really like Noctua fans, for a while they were obviously the best fans by a significant margin.
But for all their tight tolerances and exotic materials and a high price to match, they generally don't outperform BeQuiet's more regular materials but use-focused fans that are half the price. Nor are they significantly better than Arctic's general purpose fans at a quarter the price.
It'd make more sense to just buy the fan optimized for the specific common purpose (airflow or radiator) than pay double for the Noctua for a more generalized fan, but is not the best at either common use case.
Seems like these days their target audience is those who believe their marketing materials about them being the best, instead of believing the benchmark performance data.
I have used Noctua fans in computers where they worked for a decade or so, even 24x7, until an upgrade or replacement of the computer was required by other reasons than because of the fans.
I have also had many problems caused by cheaper fans.
So now I always prefer to use rather expensive fans and power supplies, from brands with which I have accumulated many years of experience, for peace of mind.
Perhaps other brands of fans that nowadays give similar results in benchmarks also have similar reliability, but I am not willing to bet on it.
If we're going by anecdotes, my last Noctuas showing signs of failure (I had 6 of them, one was ~200rpm slower than it should be, one took a several seconds longer to start spinning from a stop) about a year after the end of warranty was partially why I retired them. Same with the set of Noctuas before them (apparently my first set was from 2010). I suppose they all technically still spun so they were still usable, just not to original performance; still, hard to be too upset about the product making it through the long warranty period without issue.
But my Arctics that was installed in the same case that ran for the same amount of time are still chugging along strong, and those are about as cheap as fans get. Different load/use case though so it's probably not a fair comparison.
These days, I really think the competition has caught up or passed Noctua.
2×? Try 5× for the Noctua NF-A12x25 compared the the Arctic P12 Pro that matches or beats it in most metrics. Which isn't to say the Noctua fan is bad, it's just a luxury product for reasons other than performance.
2x more than other premium offerings that often perform noticeably better, which I'd say are usually from BeQuiet, LianLi, and Phanteks.
But yes, sometimes up to 5x more than the comparative Arctic in common size categories where it basically trades blows for most metrics that matter. Arctic is seriously unbeatable in value:performance if you just need a basic fan without other QoL or aesthetic features.
120mm is the most competitive category, and it's the most obvious category how Noctua can't keep up with the faster iterating/innovating competition.
Disclaimer: I read HWCooling like everyone serious about the subject. These reviews aren't everything, the appalling QC that results in resonances or coil whine lottery isn't mentioned.
In general, yes, Noctua is overpriced and Arctic is an incredible value, but when you want to optimize your silence/performance ratio, it's still Noctua, BeQuiet or (sometimes) Thermalright.
This was a fun revelation when I got into watercooling. You might not hear coil whine over a gpus fans. But remove the fans and put it under load and whoo boy.
So this confuses social media discussions on the topic by mixing together everyone's reports, regardless of their level of acoustic masking. "My card has no whine!" says the guy with three 2000 rpm fans going etc.
Gpu waterblocks seem to be shifting towards fully enclosed "tomb" style and I can't help but wonder if coil whine contributed to that decision.
But on topic, I had seven a12x25 in my last build, two a12 and four a20 in my current build. They are exceptional. A computer is as quiet as it's loudest part. If your care about noise, why would you ever skimp on the moving parts.
I think GPU waterblocks are becoming fully enclosed because there are so many hot components on the back of the GPU now. They were designed to rely on random case air turbulence to passively cool, but there typically isn't much airflow over the back of the card when the stock cooler is replaced with a waterblock.
Problem becomes worse when the cards are driven harder because there's more cooling capacity from the watercooling in the front, but the passive cooling capacity on the back is still the same.
I used to stick a giant fin block on the back of the card to keep temps there reasonable. I'd love it if actively cooled backplates become the norm for watercooling.
The Arctic fans are known to hum at certain speeds. This may, or may not matter to you, and certainly depends on how low the "noise floor" in workspace is.
But if they're letting me using AI for free without logging in and I just need a dumb AI slop answer, then I'm more than happy to burn their tokens instead of my own. Any serious work goes to a different LLM provider. The switching cost for moving to a different LLM provider in the future is practically zero.
reply