LesZedCB

how would they have performed if they tried to cheat but didn't find the poisoned answers? would they have had to come up with answers on their own? suppose they cheated for 10% of the answers but had to give honest effort on the remaining 90% and just made the cut?

i don't think it's very clear cut or obviously perfectly fair.

eh the highway probably has high utility value. just make it quick and painless.

some ants love to work, it brings them meaning.

some ants love to mountain bike, and who can blame them?

> Globally, natural lake volume declined at a net rate of −26.38 ± 1.59 Gt year

the phrase you read said "Over half (53 ± 2%) of the large lakes experienced significant water losses" (emphasis mine)

the fact this came from a "stochastic parrot" still blows my mind.

imagine publishing a whole paper to win an "Um, actually" argument

If the Netflix show Unstable is a reliable source, a hyperdense "biomass" or recaptured carbon cement is the future. ;) it's quite enjoyable - very silly.

services are IMO the biggest misfire of an attempted rails convention. "service" is completely domain agnostic and because there's no standard API like there is for controllers or model, the "service" API ranges from `MyService.call` to `MyService.new.call` `MyService.do_thing` etc etc.

service became the junk drawer of junk drawers even worse than lib because people at least 'believe' there is supposed to be a convention when there isn't. at least you know lib is kinda wild west.

rails has failed to channel applications into domain logic once it scales past start-up rapid development to a huge team of ICs

prescriptivists are a real bummer

i personally really like the idea of what clojure has done with spec or malli instead of type annotations. it just feels to me more like what rubyists want: pragmatic validation of "does this quack?" but like you say, i'm not sure ergonomics can be brought up to the level of ease/happiness to be used regularly

Tim Apple - 99M [1]

Satya Nadella - 54M [2]

[1] https://www1.salary.com/Tim-Cook-Salary-Bonus-Stock-Options-...

[2] https://www1.salary.com/Satya-Nadella-Salary-Bonus-Stock-Opt...

i bet we'll see more instances of companies like replika use chatbots to emotionally manipulate users into staying and paying.

what this proves to me without a doubt is that silo'd and proprietary iteration i still very clearly also a massive disadvantage. i really hope companies internalize that. if they just keep scooping up and hiding open-source improvements they very well may still be left behind.

the final quote from the doc:

> And in the end, OpenAI doesn’t matter. They are making the same mistakes we are in their posture relative to open source, and their ability to maintain an edge is necessarily in question. Open source alternatives can and will eventually eclipse them unless they change their stance. In this respect, at least, we can make the first move.

1 day of in-context learning and 1 night of fine-tuning on context. that's my pet theory, just shooting from the hip as a total layperson.

self-healing? can you describe them?

personally I think democracy is particularly fragile and requires constant work to continue reproducing.

idk i pay $30/session in the US and i don't have a diagnosis for anything, just my employer provided healthcare plan (which is admittedly quite good)

isn't it ironic that there's no autocomplete on the tool that's literally autocomplete? haha

where's the fine-tuned prompt helper model?

people feel lonely as fuck right now, and frankly i'd rather have people soothe it with an ai chatbot they parasocially love than as road rage or being a dick to the restaurant staff or many of the worse and more fatal outlets...

plus it doesn't seem to tune. no matter how many times i mark steam or fedex emails as not-spam, i just have to check there every now and again these days.

used to be cheap :'(

> Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.

https://news.ycombinator.com/newsguidelines.html

you think so? I mean, i basically stopped drinking a year or so ago because it just gave me headaches and I'm aware of the negative health affects, but I find myself in an extremely small category of people who don't drink simply because they don't want to. at least for my peers in their 30s.

I could see both synthesized being super powerful. maybe with eye tracking too?

like gestures on a trackpad

I've been a highly athletic veg/vegan for a decade and not had any major issues, sans maybe B12 or vitamin d occasionally. plenty of veg friends experienced the same. vitamin supplements are easy to take.

my local libaries (king county, WA) have dedicated meeting rooms you can book online ahead of time!

silicon valley tv show once again too accurate for comfort with its tres comas tequila.

the way i see it, the expensive part should be to train the models via simulated architectures in GPUs or TPUs or whatever.

but once they are trained, is there a way to encode the base models into hardware where inference costs are basically negligible? hopefully somebody is seeing if this is possible, using structurally encoded hardware to make inference costs basically nil/constant.

oh my god a product isn't growing, it must be garbage

employees behind a company labor to create value, they don't necessarily define the company - the owners do. owners set the policies, hours, hire and fire to promote a particular culture (or set the rules or expectations to). this power imbalance between employer and employee is exactly where the burnout tends to happen.

virtue is neither real nor transcendent.