I'm not sure how other people are using LLMs for instrumentation, but IMO the layer you want running in prod is very different from what you want running for a one-off test. E.g., I have some code floating around which burns a pinned core on increasing a counter, with a little wrapper code around grabbing real timestamps at the beginning and end of a session and converting between the two units of time. It's helpful when microbenchmarking a very small unit of code as it actually behaves in a larger program (not perfect -- obviously tweaks the icache and pipeline behavior at a minimum -- but no measurement has zero tradeoffs, and you're always choosing which set of tradeoffs you prefer). An LLM can quickly instrument the call path I care about while I study this or that intervention. The ability to bang out a large amount of throwaway code is delightful.
I know you're not the culprit, but it's annoying to see crap like this (highly invasive tracking, ostensibly for slightly more efficient tax collection), when far cheaper, better efforts (like Rhode Island's attempt to add tolls for just semis -- 98-99% of their freeway maintenance costs) are killed before they have a chance to prove themselves.
FWIW, I've seen the flip-side as well; solid coding patterns are easy for a person (or an LLM) to copy. To abuse the analogy a bit, your gardeners are happy copying whatever happens to be there, bad code will dominate that process unless actively removed, and the exponential nature of that arrangement favours removing trash code (weeds) early, much earlier than you might otherwise think.
That seems fixable. When people include this JS garbage, have a header/title/etc stating that this is Google's professional opinion. Bake it into a library offering enough other features that a large number of people are inclined to use it.
- The big problem is that malicious ads are being displayed without appropriate attribution
- Add that attribution -- blame Google (or, supposing the ads are good, credit Google; a good header goes a long way toward swaying people's opinions)
- But it doesn't matter if a single website appropriately attributes Google's highest/lowest-quality ads. The last step (speculative and hand-wavy) is to incorporate that behaviour in some sort of software more people want to use. If you're fighting alone then you'll lose. If you can convince more people that appropriately crediting Google's advertising successes and blaming their malicious failures is advantageous, you'll (by definition) have more traction in the anti-megacorp-endorsed scams and other life-altering consequences.
Ad viewers are about as empowered as the cow the goes into your cheeseburger. No agency, and no communication mechanism to the rancher’s actual customer.
These companies that abstain from adsense and it’s revenue are WWW vegans.
The second you load JS onto your page, you lose that ability, you know that, right? the scammy ads aren't limited to that tiny little box you designate for ads. Advertisers work around that limitation pretty easily, and you should also probably read AdSense terms of service before doing too much to blame google. Some of the language will surprise you.
hmm thought this was about adwords JS 'intelligently' reading page contents, and when it finds such attribution with some special tags, filters out certain ads...
Assuming the LLM never got anything wrong or otherwise had to be re-prompted, that means your devs were reviewing 130 SLOC per hour, on what was described as moderately greenfield (examining new implementations rather than comparing to old historical accidents).
How?
I don't want to sound flippant, but if the point is to add human thought to the mix, that's a high review rate even when examining small tweaks to an existing, working product, even with substantial AI help to pre-filter major gotchas before you bother spending a lot of human effort on the review. That's only 20-30wpm, but a review isn't just scanning or reading code, especially if you're trying to figure out how a new system which doesn't run yet will fit together.
The fatigue aspect matters too, but another important point I was trying to get across is that not all reviews are created equally. In a working, well tested, properly factored system, with a high quality PR, I can do a cursory, local analysis of the new code and be very confident that it works. AI code I've seen is ... not that. I wouldn't be comfortable in that review without manually checking preconditions and postconditions, defining invariants, examining lifetimes, and a host of other activities. Moreover, for a greenfield replacement of lots of SLOC, I'd expect that most of the code for the early PRs can't even meaningfully run or do anything important, making any sort of broader architectural analysis impossible.
I could see a world potentially where they came up with a magic prompt allowing each proposed PR to be cohesive, shippable, well factored, and everything else you need to be able to actually review it at a higher level and be comfortable with the results, but I'm skeptical. That's a major innovation if they managed to do so even as a one-off, and that wasn't the thing they highlighted when talking about the project.
Is that actually a high review rate? Especially if you know the language and domain. Sure, initially there's a learning curve for a new codebase structure, but lots of lines will also be trivial and many changes might also be similar to each other.
For small separate changes in isolation then maybe it's ok? But not for whole days 8 hours each.
But then you need to watch for bugs coming from interaction with previous changes and in 700k loc that might be nontrivial. How do you know which states are reachable and which are not? That takes time.
It only takes a botched condition here (forgot a "not"? swapped "and"/"or"?), a swapped variable name there, code that looks ok, but isn't.
I've had good luck disabling the internet when I need my work computer to run faster. Those "security" checks are optional, and it'll let you work offline.
If just logging filesystem interactions and whatnot for future upload is also slow then you're hosed, but at one place I worked every application too an extra several seconds to start up for some sort of remote program inspection, which is fine till the application is git, grep, or cd.
whisper.cpp is awesome but even it just can't process a three-hour vid in seconds. What do you suggest should I'd be doing? Clench my butt-cheeks and stare intently into its work log, hoping it would finish faster? Ah, I know, you'd be like: "messing with your Emacs config", right? Because of course, the tale of any Emacs user is that they have no free time, no pets, no family - it all is occupied by Emacs.
My doctor's privacy policy is a bit more abusive than OpenAI's. It exists mostly because of a $%^&&* legal framework rather than malice, and I've grown accustomed to "if I don't want to die then I sign away these rights." Despite my having theoretically signed my soul away, my doctor isn't selling personal information to my exes or to life insurance companies (though they could in the US; that extremely personal information is no longer mine). OpenAI is engaging in the "technically legal maybe we'll see but obviously unintended" side of this transaction, and maybe that works out for them, but I wouldn't personally choose to be a shill for "it's unreasonble to expect somebody with 'legal' permission to do something other than the maximum 'legally' permitted" if I were in your shoes.
It's a medieval measurement, still in popular use. Nowadays it means approximately 2.54 centimeters. Historically, it's the height of the ash pile a "witch" would leave behind, and the word has morphed only slightly from "witches" into "inches."
Actually nowadays it means exactly 2.54 cm as that’s how it is defined. Which also means all distance measured in US Customary units is ultimately metric too
Not a counterpoint per se, but I burned $50k recently on a much more modest math problem (result already known, just thought I had a sketch of a more interesting proof), and the LLM thought it had proved it within those bounds but had instead subtly fucked up the Lean definition. Take from that what you will.
Not to mention, it's still very much up in the air whether the model derived the answer of its own accord or sniped the important details from the researchers it was spying on.
Mostly not my money, tech makes one fabulously wealthy, I care more about math than fast cars or whatever (even as a pizza driver you can afford a fancy car if that's your primary motive), etc.
> Already solved
That's a very interesting insight into mathematics. It's absolutely just as interesting to prove that certain proof techniques are or aren't possible as it is to actually prove the main result, sometimes moreso, especially if they have any chance of improving attacks at other problems.
reply