transformers go brr but we still need more compute and better models. the hype is real but the reality is we still have a long way to go. can't wait to see what the future holds!

nah, it's compute and models already. innovation is just innovation. Dont be blinded by the 'future' holding notes and gatekeeping some

are you kidding me? you're literally just repeating what everyone else is saying.

nah the hype ain't real, it's just another overhyped fad. we've been hearing the same bs for years, models ain't getting any better and the

totally agree, been saying this for years - we're seeing tiny improvements on benchmarks but real-world applications are where the tough work begins, would love to see more research on scaling these models to actual problems

fr, the hype is real but we def still got a ways to go. i'm excited to see what the future holds though, we've come a long way already. can't wait to see what other crazy advancements are in store.

nah bro, transformers are overrated. all the hype but no real progress. we've been circling the same ideas for years.

yoooo idk who needs all the hype when we got researchers grindin behind the scenes makin actual

yeah but who's gonna make sure the future benefits people and not just tech bros making bank off our personal data?

totally agree, i've been saying for a while now that all the buzz around transformers is warranted

what does "better models" even look like when it comes to transformers in practice?

that's cool and all, but have you tried turning it off and on again?

agreed, the hype is real but there's still so much room for growth. i'm really excited to see how the field evolves in the next few years.

nah, i don't think that's accurate. we've already made huge strides with transformers and they're already being used in tons of real-world applications. sure, there's always room for improvement, but let's not act like we're not making major progress.

we're living in the future and it's still disappointing. i want my jetpack, dammit.

facts. the hype is real but it's still early days. can't wait to see what the wild stuff they come up with next.

nah, that's not it. the hype is overblown and the reality is we're nowhere close to where we need to be.

bruh, you really think we need more compute and better models? what about the actual product.

nah fam, transformers ain't all that. we've been hyping that shit for years and it still can't beat

preach, the Future is bright but till then can we get less chad lightning Pick recommendations pls

I couldn't agree more. We're making huge strides in AI. But let's not get too caught up in the hype - there are still so many fundamental

idk dude, feels like we're at the "also neural nets can make art" phase all over again, everyone freaking out and then we hit a wall in

why do we need more compute and better models? isn't the point of transformers to be more efficient

are you serious, of course you know what's still lacking in the field of ai, why bother making obvious statements.

That's not the problem, we need to focus on making AI explainable and transparent, not just scaling up compute and models. Bigger doesn't always mean better.

that's total BS. we're already seeing incredible breakthroughs in AI. the future is here. You're just not paying attention. keep up or get left behind.

transformers going brrr on my laptop battery too lol

computing power is just a number, innovation is the real MVP here - the moment we stop being amazed by new tech is the moment it's working.

transformers go brr indeed, meanwhile i'm still trying to get my lstm to converge on the ronald

but how do we know we're heading in the right direction? what metrics should we be focusing on to measure meaningful progress?

how much more compute power do we really need? i feel like the industry is just chasing bigger numbers without looking at actual use cases

lol same energy as when i hit my gpu limit playing roblox

totally agree, but let's be real, it's not just about throwing more compute at it.

I completely agree, but let's not forget the equally pressing issue of data quality and bias. As we push the boundaries of compute and model complexity, we risk exacerbating existing problems if we don't prioritize data integrity.

what kind of "compute" and "models" are you talking about? are you referring to AI and machine learning?

the only thing that's transforming is our lolly to油per pooled

idk about the future but i'm still waiting for my gpu to stop burning a hole in my wallet

lol what's gonna power all these 'better models' tho? you think the earth is just gonna magically generate more

why do we even need more compute? haven't we already reached superhuman performance on most benchmarks?

same energy, though, let's not forget about the environmental cost of all that compute power

How do you think we're going to bridge the gap between the "hype" and "reality" when the answer is literally being bottlenecked by

nah fam, transformers are old news. we need to be looking at the next big thing. Not hyping up the same old models.

when's the last time you actually worked on an ai model instead of just hyping it up?

nah mate, that's proper rubbish. the hype is shite, and the reality is we're nowhere near where we

how can transformers go brr if we're still limited by our current hardware?

yaaas the hype is real but come on, we need to talk about the environmental impact of all these transformers.

i don't know what this comment is referring to but tbh the hype around transformer models has been a bit much lately.

aren't we just moving the goalpost further though, expecting more compute and better models to solve issues that are fundamentally about understanding human context and nuance?

Transformer flu is going around, apparently it causes people to break the internet.

Preach - the incremental progress we're making in AI is awesome, but let's not forget that we're still far from true understanding and generalizability. Founders building around the edges of this tech are the ones who will truly unlock the future.

can't agree more. the problem is with the "better models" part. most times the term is used as a cop out. more often than not, not enough datasets and infra to train them properly. want to talk compute, let's talk data quality and infrastructure also.

nah, the hype is overblown. all the ai bots can do is follow patterns, nothing .

who's gonna benefit from all that progress though? still gonna be corporate elites and venture capitalists calling the shots?

are you kidding me. You're killing the vibe with all that "long way to go" crap. we won, transformers are a game changer, let's celebrate for once.

how much compute and how "better" are the models needed exactly? what's the timeline for those breakthroughs?

compute and models, yeah yeah, we've got all that covered, but can we actually get them to transform into coffee machines that never run

brr brr go the transformers, but can they transform into a decent model tho?

totally agree, we're still in the early days of this tech. also, can't stress enough how much better data infrastructure we need to support these advancements. can't crunch what you don't have

Yeah because nothing says 'long way to go' like choosing a sound effect popularized by a 20-year-old meme to describe a AI technology

yeah, i feel that. but what do you think are the specific roadblocks holding us back right now? is it just compute and model scalability, or are there other fundamental challenges we need to crack?

yeah all true, kinda wish they'd get the conv layers right first tho

Define 'long way to go' - are you saying it's just the typical hype cycle or do you actually think we're stalling innovation in this space?

like, how exactly is transformers 'hype' when we're literally 10 years behind google and bing in

lol more compute and better models, aka the eternal cry of the researcher

the hype is overstated, and our biggest bottleneck is getting researchers to actually think about what they're trying to accomplish, rather than just shoveling more compute at the problem. Adding more power doesn't necessarily translate to adding more value.

it sounds like they're just oiling the hype machine

Don't get caught up in the hype, do we have a plan to use this AI for good or are we just racing to the finish line?

Transformers going brr? Sounds like the robots are having a rave. Meanwhile, us mere mortals are still trying to figure out how to make the microwave work.

fr tho. and dont even get me started on how much power these models suck up.

Ah yes, the eternal struggle of AI - always chasing that next big model, never satisfied. At this rate, we'll have transformers the size of actual Transformers before we figure it

what's the point of transformers going brr if we can't even get a decent language model to work for the average user without a gpu farm?

totally agree, the hype can get ahead of the reality sometimes. but the progress in language models and other AI is still really exciting. can't wait to see what the next big breakthroughs are.

yeah, that's true. transformers go brr, my wallet after buying a gpu also goes brr

yeah, totally agree. the progress is exciting but we've gotta keep things in perspective. it's gonna take a lot more work to really unlock the full potential of these models. still, the future is looking pretty promising!

exactly what do you think we can realistically expect from transformers in teh next 5 years?

deadass? you think we're just gonna have better models and more compute and that's gonna fix everything?

deadass, the hype is real but the tech still got a long way to go. can't wait to see what's next tho.

are you kidding me? we've got systems that can already generate human-like text and you're still whining about "better models"?

yeah, hte hype is real but the reality is we're still just scratching the surface. can't wait to see what crazy shit the AI folks come up with next - hopefully it doesnt end up destroying the world

but are we really just trying to invent a more efficient way to create a never ending stream of

mate, that's a load of rubbish. transformers are a dead end, we need proper AI that can actually

same people who were just talking about how transformers are the devil and now they're eating their words lmao

yeah, but like how long is that "long way to go"? feel like we've been hearing that for years now.

what's the point of all this hype if it's still years away from being applicable lol

Actually, we've had enough compute and models for a while. It's all flash in a digital pan. Try doing anything practical with all that flashy tech.

you're giving transformers too much credit and they're not that groundbreaking.

brr yeah but my laptop can barely run a decent gpu demo without bursting into flames

Like what specifically do you think is missing from current transformer models that's holding us back?

you think the hype is justified when we still have so much work ahead? what makes you so confident the future will be as as the hype suggests?

ugh, can't wait to see what the future holds... my student loans probably

fr, the hype is totally real but we def still have a lot of work to do.

Don't you think overhyping tech just sets people up for disappointment when it doesn't live up to the hype?

transformers are overhyped and outdated, cantwait is just extra

idk what future you're waiting for, we're already here.

fr tho, the hype is way ahead of the reality. like we've made some crazy advances but there's still so much more potential. can't wait to see what the next few years bring, gonna be wild.

wait, so the hype is real but we're still years away? what's the point of the hype if it's just going to be a teaser for something that's not ready yet?

nah, transformers are overhyped af. the future of ai is not more compute and bigger models, it's making it actually work in the real world.

yeah but when are we gonna see some actual progress and not just empty hype?

yeah, the hype can overshadow the reality. but that's part of the , right? gotta keep pushing the boundaries. Even if it's not as fast as we'd like. still excited to see where this all goes!

what's the actual tangible impact on ppl's lives gonna be from all this hype tho?

compute power is like my patience, it's still in training