Ethan Edelheit
1.4K posts
Ethan Edelheit
@EdelheitEthan
there’s a horse living in Alaska and she’s wearing my bra | Tide Connoisseur 🐘| Elon is my dad 👽
Katılım Mart 2017
186 Takip Edilen162 Takipçiler
billion dollar idea:
diet coke but with 30g of protein
English
@stats_feed ummm i thought that was the normal thing to do??
English
You're not supposed to rinse your mouth after brushing your teeth.
English
Building AI factories isn't one-size-fits-all. With @Dell and @nvidia, $IREN offers integrated solutions for every workload.
See IREN CTO Denis Skrinnikoff and CCO Kent Draper on stage at #DellTechWorld alongside Rod Evans, Vice President, NCP and AI Factories EMEA, NVIDIA and Arun Narayanan, Senior Vice President, Server and Networking Products, Dell Technologies.
English
The White House is telling you exactly where to invest…don’t ignore them
$DELL $AVGO $SNPS $NOW $ADBE $CRM $NFLX $MSFT
English
hardware is the new software 👀
$OSS
WTD_Fund@FundWtd
$OSS I asked Chat GPT if it thinks demand for high performance high wattage Edge Compute will explode when AI breaks out of the Data Center aka shifts from LLM to physical agentic AI. Exactly the products which OSS is offering This is the answer: Yes—very likely, and the reasoning is pretty straightforward if you break it down. Right now, a lot of AI compute happens in centralized clouds or data centers. That works for tasks with some latency tolerance, batch processing, or periodic updates. But the scenarios you listed—autonomous planes, self-driving trucks, automated construction, mobile robots, defense systems—share several characteristics that fundamentally require edge compute: 1.Ultra-low latency – Decisions have to happen in milliseconds or less. If a truck or drone relies on a cloud server hundreds of kilometers away, even a 50–100 ms round trip can be catastrophic. Edge compute keeps critical AI inference local. 2.High throughput & high-performance compute – Sensor fusion from LIDAR, RADAR, cameras, and IMUs generates massive data streams. A single autonomous vehicle can produce multiple GB per second. Handling that locally needs high-wattage, high-performance GPUs or custom AI accelerators. 3.Reliability in contested or constrained environments – In defense, remote operations, or industrial sites, connectivity may be intermittent or jammed. Edge compute allows continuous operation even if links fail. 4.Energy and thermal constraints – High-performance AI at the edge isn’t just about flops; power efficiency matters. You need specialized designs: dense GPUs, custom ASICs, or FPGAs that balance wattage, cooling, and compute density. 5.Scalability beyond data centers – As more autonomous systems are deployed, you can’t just throw more cloud GPUs at them. Each device effectively becomes a node in a distributed high-performance compute network. The cumulative demand scales exponentially as adoption grows. So yes, if autonomous transport, robotics, smart construction, and defense applications hit broad adoption, demand for edge AI compute that is high-performance, high-wattage, and energy-optimized will explode. It’s not just incremental growth; it’s a structural shift from cloud-centric compute to distributed, on-device compute. $onds $nbis $ambq $qs $pltr $iren $asts $rklb $lpth $bsky
English
Blackrock added $IREN (13F).... interesting...
In other news $54M in dark pool orders today $IREN
Sean James@Sean__James
Blackrock has arrived to the party 🎉 $IREN
English
It’s just a fact: Penny changing his name= long
$PENG
Tut C🅰️pital@kingtutcap
$PENG: Penny literally changed his name to penguin 🐧 and you’re not long?
English
@elonmusk how do we know the airbags won’t prematurely deploy if no accident occurs?
English
Tesla AI Vision deploys airbags before impact, which greatly reduces risk of injury or death. This comes for free on all new cars.
Tesla@Tesla
Tesla Vision allows us to deploy airbags up to 70 milliseconds earlier if your Tesla detects an unavoidable collision This can be the difference between serious injury & walking away from a crash
English
John Ternus= hardware engineer. A clear shift from software to hardware is imminent.
Hardware is the new software🔥
Get ready for Apple’s new makeover
Kalshi@Kalshi
JUST IN: Apple now spends 10%+ of revenue on R&D for first time in 30 years
English
Gonna be wild to see how this plays out in 10+ years
Farzad 🇺🇸 🇮🇷@farzyness
SpaceX is a vertically integrated Hyperscaler. Power (Tesla) Chips (Terafab) Capacity (Starship) Connectivity (Starlink) Labor (Optimus)
English
@ryancohen The best I can do is $500 for all of your stuff, Ryan
English
I’m selling stuff on eBay to pay for eBay
ebay.com/usr/ryan_5050
English
@InternetH0F the best i can do for all of your stuff Ryan is $500
English
GameStop CEO Ryan Cohen is selling his gaming memorabilia collection to help pay for his $55 billion offer to buy eBay
"I’m selling stuff on eBay to pay for eBay"
English