GreenCloud | #1 TOP PROVIDER 2024 SALE | DOUBLE PROMOTIONS with GENOA/9950x | +$2000 GIVEAWAYS

JohnnySac · February 2025

@FAT32 said:

@nullnothere said:

@FAT32 said:
What I like about having industry experience is that I overthink and overengineer the heck out of too many shit worrying it cannot handle the high traffic leads to downtime

Super-Over-Engineered-MJJ-Proof-V^32-IP-Coupon-Protector-With-NO-DEALS-LOST when?

I said that because I have so many ideas in mind kinda blocked by my own overthinking and over-engineering. But looking at the fun project by @beanman109 always makes me feel like those were the day I just spend short amount of time to write some codes and make some pages without worrying how to scale.

(Well premature optimisation is root of all evil... easier said than done)

Totally get this!

undoIT · February 2025

I think I'm done with GreenCloud deals. The banter on this board has been a blast but I wasted way too much time trying to get deals and have grown tired of being cock-blocked after doing everything right on my end. Hate to leave the thread on a sour note, but that's I'm feeling at the moment.

cybertech · February 2025

@undoIT said:
I think I'm done with GreenCloud deals. The banter on this board has been a blast but I wasted way too much time trying to get deals and have grown tired of being cock-blocked after doing everything right on my end. Hate to leave the thread on a sour note, but that's I'm feeling at the moment.

next time then good luck

FAT32 · February 2025

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

karanchoo · February 2025

@cybertech said: @undoIT said:

I think I'm done with GreenCloud deals. The banter on this board has been a blast but I wasted way too much time trying to get deals and have grown tired of being cock-blocked after doing everything right on my end. Hate to leave the thread on a sour note, but that's I'm feeling at the moment.

same here.

nullnothere · February 2025

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

Point noted - I'm getting old.

cybertech · February 2025

ya know after all this the first VPS im most excited to put to use is actually the $7.75 9950X

i feel like a boy with his new toys

beanman109 · February 2025

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

FAT32 · February 2025

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

I am thinking more of attack scenario - Like can the service easily handles thousands of QPS without any issues? Is there any potential security issues? Can a specific part be made high availability? Is there any redundancy in the design? Can the SPOF be mitigated?

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

emgh · February 2025

@FAT32 said:

@emgh said:

@pkr said:
Please don't post nude photos. Cannot open this thread if someone is around.

sorry I'll continue

I would prefer if you being a little bit more nice to anyone in the thread...

Sorry daddy I’ll behave

komdragon · February 2025

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

emgh · February 2025

@komdragon said:

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

Isn’t it confirmed that DeepSeek was trained using the OpenAI API? If so, not surprising they didn’t need lots of GPU..

nullnothere · February 2025

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

I am thinking more of attack scenario - Like can the service easily handles thousands of QPS without any issues? Is there any potential security issues? Can a specific part be made high availability? Is there any redundancy in the design? Can the SPOF be mitigated?

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

Can't resist (can easily be extrapolated to over engineering I suppose):

cybertech · February 2025

man its getting hard to remember exactly what's what

komdragon · February 2025

@emgh said:

@komdragon said:

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

Isn’t it confirmed that DeepSeek was trained using the OpenAI API? If so, not surprising they didn’t need lots of GPU..

They did use that too, but its design is more efficient too. DeepSeek also definitely lied about many things but its architecture is more efficient without a doubt.

I don't think that OpenAI has any right to complain though, given they trained their models on everyone's content without asking or paying.

NaughtyWombat · February 2025

725

geo · February 2025

komdragon · February 2025

@geo said:

That "real data" is content taken from everyone without permission.

emgh · February 2025

@komdragon said:

@emgh said:

@komdragon said:

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

Isn’t it confirmed that DeepSeek was trained using the OpenAI API? If so, not surprising they didn’t need lots of GPU..

They did use that too, but its design is more efficient too. DeepSeek also definitely lied about many things but its architecture is more efficient without a doubt.

I don't think that OpenAI has any right to complain though, given they trained their models on everyone's content without asking or paying.

For sure. Zero rights to complain.

allthemtings · February 2025

@cybertech said:
man its getting hard to remember exactly what's what

juicy

nullnothere · February 2025

@komdragon said:

@emgh said:

@komdragon said:

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

Isn’t it confirmed that DeepSeek was trained using the OpenAI API? If so, not surprising they didn’t need lots of GPU..

They did use that too, but its design is more efficient too.

Relevant (?) read:

https://semianalysis.com/2025/01/31/deepseek-debates/

Just in case it helps.

I have no clear way to judge either way.

CharityHost_org · February 2025

@barbaros said: @emgh CockCharity when or CharityCock? I'm just too sleepy to come up with creative names.

foul foul foul. Truth bites all sorts of ways. Dont sign up or call on @CharityHost_org if you dont like truth and reality.

NHNHNH000 · February 2025

CharityHost_org · February 2025

@Saragoldfarb said:

Romania. Does it matter where the location is even? Romania has good networking, but we are no where close to that or even trying to with that. Go Romania hosts!

komdragon · February 2025

@nullnothere said:

@komdragon said:

@emgh said:

@komdragon said:

@FAT32 said:

@beanman109 said:

@FAT32 said:

@nullnothere said: I can relate. As I've grown older (shush, don't ask embarassing questions, we're in a polite forum), I've realized many times that what we know often comes in the way of doing things and it's not easy to break free from the scaffolding/limits of that knowledge. I think that's why fresh, young minds who are not limited by things like pick-any-2-of-3 type of frameworks can come up with some really innovative ways of solving problems, which is often times remarkably beautiful AFTER you see the solution (and wonder why YOU didn't think of it).

I don't think it is the kind where I didn't think of it. But being in the industry with hundred millions of users means that I have seen a lot of different types of scaling issues, which leads me to be very careful when planning the architecture.

wouldn't a tickbox that asks "Are you an MJJ?" suffice?
if you're worried about people lying when ticking the box then just add under it
"Did you lie on the above question?"

But honestly, with computing resources getting cheaper and cheaper, I might just be thinking too much

This is the thinking that allowed DeepSeek to get ahead of ChatGPT/others, at least in some ways. China having limited GPU access forced them to be more efficient, while US companies had no issues getting truckloads after truckloads of GPUs so they did not bother to optimize.

Isn’t it confirmed that DeepSeek was trained using the OpenAI API? If so, not surprising they didn’t need lots of GPU..

They did use that too, but its design is more efficient too.

Relevant (?) read:

https://semianalysis.com/2025/01/31/deepseek-debates/

Just in case it helps.

I have no clear way to judge either way.

You can actually run DeepSeek locally on much lower hardware than competitors, so there's that.

CharityHost_org · February 2025

@emgh said: CharityCock 100 %

And I dont have to wonder what kind of person you are. Just pure nasty banter and abuse.

emgh · February 2025

@CharityHost_org said:

@barbaros said: @emgh CockCharity when or CharityCock? I'm just too sleepy to come up with creative names.

foul foul foul. Truth bites all sorts of ways. Dont sign up or call on @CharityHost_org if you dont like truth and reality.

Brudda are you god?

allthemtings · February 2025

emgh · February 2025

@CharityHost_org said:

@Saragoldfarb said:

Romania. Does it matter where the location is even? Romania has good networking, but we are no where close to that or even trying to with that. Go Romania hosts!

My $2.99 deal’s in the US so you may match that

emgh · February 2025

@CharityHost_org said:

@emgh said: CharityCock 100 %

And I dont have to wonder what kind of person you are. Just pure nasty banter and abuse.

[removed to not trigger and start a discussion]

Howdy, Stranger!

Categories

In this Discussion

GreenCloud | #1 TOP PROVIDER 2024 SALE | DOUBLE PROMOTIONS with GENOA/9950x | +$2000 GIVEAWAYS

Comments

725

Howdy, Stranger!

Quick Links

Categories

In this Discussion

GreenCloud | #1 TOP PROVIDER 2024 SALE | DOUBLE PROMOTIONS with GENOA/9950x | +$2000 GIVEAWAYS

Comments

725