Yes my name is Ulric, no I'm not German
Joined February 2010
- Tweets137
- Following143
- Followers28
- Likes127
U.SOL retweeted
WARNING: Vitalik says if crypto keeps centering on gambling with no real-world use, the industry will die fast.
U.SOL retweeted
Sup, you degenerate fucks. Timmy's GIVEAWAY ALERT! 🚨
I've been cooking heavily lately and thought I'd reward y'all since most of you are probably sidelined
I'm dropping $75 in $SOL to ONE lucky degen to fuel your next gamble!
Here are the rules:
- Follow @TurnDegen
- Like
- RT
- Bookmark
- Reply with your wildest trade story
Winner get picked in 24H
U.SOL retweeted
“So you bought $ETH at $4,950 because Tom Lee said it was going to $30,000?"
“Yes, Dave.”
“Then you sold it at $2,600 because the bears convinced you it was over?"
“That is correct, Dave.”
U.SOL retweeted
🇺🇸 NVIDIA VS. GOOGLE: THE AI CHIP WAR JUST WENT FROM “HEATED” TO “BOSS FIGHT”
Nvidia took a 3% punch to the face after reports Meta might cozy up with Google’s TPUs - and Jensen Huang responded:
“We’re a generation ahead of the entire industry.”
In other words: Don’t get it twisted - they’re good, we’re god-tier.
Google’s TPUs have been getting hype ever since Gemini 3 dropped, trained entirely on Google’s own silicon.
It’s the closest thing NVDA has seen to a real challenger in years - not a GPU replacement, but a “what if the in-house guy actually showed up shredded” moment.
But here’s the read:
TPUs are a Ferrari. Nvidia is the highway.
Google doesn’t even sell their chips - they rent them.
Nvidia sells to everyone, supports everything, and scales like a religion.
Blackwell GPUs are still the industry’s gold standard, and Google knows it - DeepMind’s Demis Hassabis literally texted Jensen to say the scaling laws still hold.
You can practically hear Huang smiling.
This isn’t a dethroning. It’s the first real market pressure Nvidia’s felt since AI went mainstream - and Jensen will answer with a new architecture drop that makes Blackwell look mid.
Meta? They’ll dual-wield: TPUs for training runs, Nvidia for everything else.
The AI arms race just hit Act II. And nobody’s playing nice anymore.
Source: CNBC
🚨 THE TPU ERA: GOOGLE’S SECRET WEAPON ISN’T “COOLER TECH” - IT’S A CHEAPER AI FUTURE
Google’s TPU story didn’t start as a flex. It started as a panic attack.
Back in 2013, Google realized that if everyone used voice search for 3 minutes a day, they’d have to double their data centers.
CPUs and GPUs couldn’t handle it without lighting cash on fire.
So Google built its own chip - not for gamers, not for crypto bros, but to stop its own AI from bankrupting the company.
Here’s the simplest version:
GPUs = Swiss Army knives.
TPUs = scalpels.
Nvidia’s GPUs do everything - which is why everyone uses them.
Google’s TPUs do a few specific AI tasks insanely well - and insanely cheap.
The punchline everyone cares about?
TPUs often beat Nvidia on cost, speed, and efficiency for the right workloads.
Sometimes 30% better… sometimes double.
The catch?
TPUs are locked inside Google Cloud. CUDA is the industry religion. And companies don’t want to rebuild their whole codebase just to save money.
But here’s the part Wall Street should be sweating:
If Google keeps scaling TPUs - and clients stop caring about training on GPUs - GCP becomes the only major cloud that isn’t forced to live under Jensen Huang’s 75% GPU tax.
Prediction?
Within 2 years, TPUs become Google Cloud’s biggest competitive weapon.
Not because they’re faster.
Because they’re cheaper, at industrial scale - and every cloud provider is about to feel the margin squeeze.
Nvidia still wins the hardware war.
But Google?
They might win the economics.
Source: @ZeroHedge
