QuantumNoodle

So this wallet thing is an identity, so when I have to use it on various sites then they effectively don't need cookies anymore to track me?

They don't want models to become commodities. Then people won't pay them.

Oh cry me a river. You went full tilt ahead scrambling before safeguards were set up and now want to set up safeguards protect your models from commoditization. Now that you've gotten your footing you preach about the dangers of powerful AI models.. Cats out of the bag. The fu*k you think was going to happen when you were building these tools? A Utopia?

Some of the code bases I have at work are legacy spaghetti -- hard to reason about. I too employ the prompt and pray strategy and then see if CI fails. I can only be surgical where I already have a grasp on the codebase, or if I was involved from the beginning.

Morally, I dislike what AI is doing to the profession. Rigor is suffering.

It really depends on th SDLC of the codebase. I've sat in the park with a pad and pen and wrote down ideas for a small-ish utility that had both a DB and some low level systems interaction. I designed all the interfaces, control flow, and user facing ui. A decade ago I was designing class diagrams with public and private methods, then coding it up. Today I can articulate it with loose words and have it come to life.

The secrete ingredient here is... Im an experienced programmer, I know what to ask and what to look out for. I know sockets and how devices work under the hood and which knobs are available for me to turn. Juniors just don't have enough experience. The question should really be "what do inspiring programmers do?"

The thing is, AI is a productivity boost but staff drive all the productivity. The catch 22 is AI is expensive so in order to provide it to everyone they need to reduce staffing. The result is remaining folks are able to produce more but the organization as a whole is not exceeding pre-layoff output. And now things are being dropped on the floor and falling in between the cracks -- which impedes efficiency and velocity. Happening at my company now.

AI inference needs to get cheaper or there will always be this "terminal velocity." I suspect AI labs' incentives to reduce costs is only where there is overlap to free up hardware/utilization (to then provide inference to more paying customers.) I highly doubt they will want to make things cheaper for users -- they have debts to pay.

Open weight models are the way and forward, it is the only way an organization can truly control costs by self hosting or buying cheaper inference. Relying on closed weight models is a business risk. Kimi models are only marginally worse than opus but significantly better than the bleeding edge of yesterday's sonnet.

For my team to re-work fragile, complex parts of the system we need to push back at product from shipping "features." Product does not see reliability as a feature they can advertise -- it's expected. And frankly I can't be bothered to keep taking side quests to fix things that increase my workload, I will burn out again. I want to fix things. But our planning revolves around implementing features customers are waiting for (aka compete in the market.) I am responsible with AI tools but not everyone I work with are. Even though responsible AI users are in the majority, our system is only as stable as the weakest link.

wHy DOnT tHeY ASk aI

$2k or $40k? One of those is not "self host."

Your TB stories made me recall my (fond) TB stories. I came from a country that requires tuberculosis vaccines as a school-entry requirement. I have the vaccine and antibodies. Then moved to a country that didn't have this requirement but I had to be tested to make sure I didn't have TB for things like camps, college, etc. The test is something like a vaccine that injects only dead TB cells, if the injection site welts up from anti-bodies then you had to get another whole panel of tests (like X-ray of lungs). Thankfully I've never had it but TB is apparently no joke. Though blasting my chest with radiation is no healthier :p

Anyway, thanks for sharing!

Agreed. Especially since there is similar functionality for explicit tracks

Great concept but I did not have a ton of fun playing after the novelty wore off after a few minutes. It would've been more fun if time stood still and I had the opportunity to plan what I do at each cpu cycle. I was looking forward to managing cpu cache hits and ram usage.

ANES? Like the homophone to anus?

Good point. I've (even with agents) never made more than like 5 PRs in one day internal to a company and if I have they typically included accompanying proto or submodule changes. Heck give a factor of safety of 2x and cap at 10 daily PRs per account for repos that youre "untrusted"

I get a lot of nostalgia and initial interest in these types of phones. But I always talk myself out of it because, with self control, my current phone is just as good plus it's one I already have. Perhaps when my current phone loses support then I'll consider these alternatives. But for the price... Eek. Rather just buy a "real" flip phone that can give out a hotspot.

_dont create benchmarks that will incentivize ai labs to optimize towards... Especially ones like battle royal!_

lmao excellent reference! While this is poking fun at the block chain hype a bit, a shocking amount of (good) software engs don't know good security practices. I'd be interested to hear arguments for and against being licensed to write software for particular applications (e.g J.Doe can work on this application but any auth code has to be signed off by someone licensed). AI tools are great but screwing up an implementation (for a well thought out spec) does happen, esp. in security applications and someone needs to be liable for life-ruining data breaches.

Given how many critical systems software touches these days, I am surprised that there they are not licensed. Imagine if civil engineers can just go building shit as minimum viable products? Sure prototype quickly but someone should be found liable for final product.

Do you have any resources on hardware necessary for running models and tweaks? I see you mention 2x 3090 and I wanted to do more search on what hardware is satisfactory for what models.

I am in no way supportive of monopolies though taking "the need to compete" out of the equation and having steady flow of money that won't be cut off if "someone beats you to market" was pretty revolutionary. Consider Bell Labs, staffed with real "nerds" that never had to prostitute themselves or to pitch an idea on the merits of making money -- they simply did stuff that was fun, in ther nerdy way. Because of this we have wireless communication, transistors, digital cameras sensors, c, and many others.

(There were a lot of negatives due to this AT&T monopoly but we are talking about nerds here and having to socialize your own worth/value. It's a shitty game that real nerds aren't necessary interested in playing)

Man, I never hear good security things about npm

I mean, I get it. I have a family, financial obligations, etc. Certainly not shaming anyone for fulfilling those obligations working for a successful company. But don't put yourself up on a pedestal and cloak yourself in righteousness behind a veil of strong morals

I mean, I get it. I have a family, financial obligations, etc. Certainly not shaming anyone for fulfilling those obligations working for a successful company. But don't put yourself up on a pedestal and cloak yourself in righteousness behind a veil of strong morals.

I want to see more of "I'm rejecting this offer to work for Google bc my moral compass" not from folks who are paid out and ready for the next thing already.

While EU is actively trying to erode encryption and privacy of all digital communications.

Looks like he updated his "About me" section

> At the moment, my main focus is [...] fighting against (governmental or corporate) mass surveillance

While he always had "Ethics in Computer Science” as an interest, I wonder what blinds people into accepting offers at Google -- the advertising company. I want to take his words as sincere, but Google has been privacy violating for longer than his tenure with android. Money and prestige is a hell of a drug I suppose. He could very well work for grapheneos but no money or immediate persteige there (sadly).

I've worked roles where our priorities shift with the wind. Many times it is for good reason, like a strategic customer to get a foothold in a market. Other times it is just because management hyped up some effort. All's this to say, nod saying you will do it then just go about your day doing focusing on the actual priorities. Don't let workload mount up bc deadlines are all made up.

Ironic that the other trending HN post is "EU-banned pesticides found in rice, tea and spices"

https://news.ycombinator.com/item?id=48447062

Goodness, what an unreasonably long description for an issue.