- stephen_cagleFair enough. I guess I am making a bit of a straw-man in that I feel I just don't buy the idea that doing the same thing 10 times over the course of 10 years is somehow worse than doing different things over the course of 10 years. They are signals, and depending on what we are attempting, they just mean different expected outcomes. One isn't necessarily worse than another, but in this case it seems to be implying it is the distinction between Midlevel and Senior.
- You know, this is kind of a funny take at some level. Like, for any surgery, you want the doctor who has done the same operation 10 times, not the one who has 10 years of "many hat doctoring" experience.
I'm not really arguing anything here, but it is interesting that we value breadth over (hopefully) depth/mastery of a specific thing in regards to what we view as "Senior" in software.
- I honestly can't tell if you are speaking in metaphor or literally?
- And yet... is it? Realtime means real discussion, and opportunity to align ever so slightly on a common standard (which we should write down!), and an opportunity to share tacit knowledge.
It also increases the coverage area of code that each developer is at least somewhat familiar with.
On a side note, I would love if the default was for these code reviews to be recorded. That way 2 years later when I am asked to modify some module that no one has touched in that span, I could at least watch the code review and gleem something about how/why this was architect-ed the way it was.
- The email I received from them this morning claims that this will be cheaper for 96% of users...
I have cron jobs on several github projects that runs once a day and I have never been charged anything for it (other than my github membership). Should I expect to be charged for this?
- Damn, I just looked up how long I have been paying using https://payments.google.com/ . Looks like I've been paying for youtube music since October 2014. These grandfathered people must be really really early. :]
- Wait a minute, why is mine $13.99 a month?
But agree, totally worth it if you at all value your time.
- I'm curious, do they have to be? Would it be possible to boot the program + the anti cheat into it's own VM or something? So they know I am running on trusted hardware, but I know that the aren't reading my emails? Genuinely curious and don't know the answer to this.
- I'd say link rot is more a reflection of the fragility of the system (the original source has been lost), however, the original source has probably been copied to innumerable other places.
tldr: both of these things can be true.
- I think it is unfortunate how many resources are put into making things secure with TPM's and how little resource is put into basically having secure and simple sandboxing...
All I really want is a computer that allows me to fully control the permissions and filesystem access of all the programs that I manually install on my system. Almost every program (in my case) needs 0 filesystem access outside of what it installed itself and shouldn't be looking or snooping at anything that isn't in its own process space.
I want a clear and simple way to limit the blast radius of how badly a program could actually screw up my system or have access to my files.
I recently experienced the opposite of this on Android, where I tried to install a very well reviewed ebook reader called MoonReader. But MoonReader seems to require complete access to every file on my Android device to work correctly. That is insane. I looked it up a bit more and it seems that Google has simplified (or something) permissions, but now there isn't much choice other than asking for full file access (I just want to give it access to one directory).
Anywho, just a minor vent, that we are insisting that the only way to make things secure is this sort of attestation path, but we don't spend any energy just making it possible to limit the blast radius of software on most OS'.
- I feel like this review needs a huge amount of context on PTA and his previous films to create any sort of justification for this film. That... that does not seem like a good movie to me.
- I... i actually wouldn't be opposed to this if it meant a return to my buying things from the website of the company that makes the product.
What Amazon really nailed, and the reason I buy like 90% of things from them, is the easy and known shipping and return policy.
But the second reason I buy from Amazon is laziness. I am too lazy to look up the (often confusing) product on their company store, I'd rather just buy that product through amazon. To compound my laziness, I don't like having to once again fill out my email, home address, and credit card for the thousandth time when I go to their site.
But if I don't have to find and navigate to their site and I don't have to fill in my information for the thousandth time. Now, we are talking about me realistically considering a direct purchase.
This is likely good for the consumer, and largely good for the world (less centralization).
- Referenced Article in this post -> http://www.incompleteideas.net/IncIdeas/BitterLesson.html
"We want AI agents that can discover like we can, not which contain what we have discovered. Building in our discoveries only makes it harder to see how the discovering process can be done." - I am curious if people would read this as an advocacy or criticism of LLMs?
- I still can't understand how functioning adults believe that releasing their work in two separate places is a good idea (Ai Studio and Vertex AI).
- I'll be honest that I read a different article on the same topic and did not know about the salary floor. So I wasn't thinking about that. I'm... mixed on that, but it does add a wrinkle to the equation.
I prefer a purely compensation relative approach because it let's the market decide what the actual salary for a software engineer is (with a percent of compensation premium for a foreigner and a 0 extra cost for a native). The market can dynamically adjust what a software engineer makes (not fixed price control) but it just cost more to hire foreign people.
In direct response to your first sentence, I think even foreign workers (who largely work harder and have more on the line than domestic workers) would question the wisdom of working for 50k a year as a software engineer in the US. They are actors in this system as well, and you can't just assume that you could offer 50k and get them to accept.
- This straight 100k to the top is not a good way to implement this. It should be a percentage (say 50%, we can talk about what the number should actually be) of the total compensation that is being paid to the H1B. We should also just completely remove caps on H1B.
This allows companies that truly want extraordinary talent to pay a premium to acquire it with no red tape . It also makes it far less likely that they can significantly underpay foreign workers to work in the united states and undercut American employees (at a 50% surcharge, you would have to pay 2/3 the prevailing salary to break even (assuming all employees are the same)).
The 50% number is something I made up, I think we can have an honest discussion about what that number should realistically be (and it should probably be different for different industries). But my main point is it should simply be a percentage tax paid on top of all compensation for foreign employees. This is the correct way to balance domestic companies undercutting domestic labor, while allowing them to access genuinely extraordinary talent with no impedance.
- Damn, not even a month after getting a butterfly kiss of a slap on the wrist for abusing their monopoly position... and they are already pulling this?
Thank god we have strong regulation in the US to protect us. /s
- I still think it is interesting. This bubble (if you believe it) seems closer to tulips than it does to rail or real estate. People were investing in tulips with the possiblity that the market would completely crash, and tulips, without the frenzy behind them, are just tulips (not that useful). rails and real estate may loose value, but there was never any real possibilty that they would become valueless. GPU's are interesting in that they are like tulips (a breakthrough in efficiency could render them basically worthless), but they literally have the worthless aspect built into them directly. You will not be able to profitably run an LLM on a 5 year old GPU, as your competitors will be able to run inference at much higher efficiencies than you with modern chips and will undercut you on price.
It is just different because it is an almost guaranteed point in the future that they become worthless (unlike rail or real estate).
- I do wonder what a random dip in quality causes in a long running conversation? Does the conversation recover at a later point, or does the introduction of temporary idiocy permanently affect the rest of the conversation?
Statistically, probably likely that the dip occurred at a point that wasn't too important? But what happens if the idiot comes out at a critical point?
Kind of reminds me of the two alternate ways that time travel works in sci-fi. Does the small change to the past explode like a fission reaction, or does history heal itself?
Anywho, if errors do accumulate, I can see being very pissed off even with temporary idiocy from the model, as it means it poisons the context for the entire rest of the conversation.
- I mean if you have a better idea for how to assign your attention, then I am all ears. :]
I'd say trust is a pretty reasonable way to assign attention.
I guess the fairest way might theoretically be to require everything to be submitted anonymously, with maybe authorship (maybe submissionship) only being revealed after some assigned period?
This is better for the incubants, but would require a huge amount of energy compared to "Oh, simon finds this interesting, I'll take a looksy".