인포마켓 강용운
@yongwoonkang
Followers
21
Following
64
Media
136
Statuses
504
Semiconductors, AI, and real tech — clear analysis from a former capital markets reporter and derivatives trader, delivered through my YouTube channel.
Joined August 2025
Ex‑Google X chief @MGawdat. says talk of AI creating jobs is “100% crap,” warning that displacement could reach even CEOs. I agree, I'm struggling to see many new jobs in a post "AGI" world. I can see, however, many new *hobbies* emerging. In the medium term I see AI Trainers
💼 The idea that artificial intelligence will create jobs is “100% crap,” - ex-Google ex exec Mo Gawdat. He says even executive roles face automation risk as toolchains take over more work. He gives some examples, an app at his startup was built by 3 people using AI tools,
44
40
293
TSMC Begins In-House EUV Pellicle Development, May Delay High-NA EUV Adoption…Negative for ASML - TSMC’s 6-inch wafer fab (Fab 2) in Hsinchu Science Park will cease production on related lines. - According to semiconductor equipment industry sources, the 6-inch fab will be
9
23
132
After a year of scientific scrutiny, a rock sample collected by the Perseverance rover has been confirmed to contain a potential biosignature. The sample is the best candidate so far to provide evidence of ancient microbial life on Mars. https://t.co/0BAO1dhMG8
549
6K
23K
🎨 Fragmented 🎨 ChatGPT + nano-banana is a great combination! Prompt : close-up image capturing a fragmented portrait of a [SUBJECT] through [4 or 5] irregularly shaped glass shards scattered on a textured off-white background. Each shard reveals intricate details with light
14
22
205
✅ Tesla AI5 & AI6 칩 설계 요약 및 시사점 (공식 발언 + 인터뷰 기반 정리) - by LABX 부제 : Tesla AI5·AI6, 수직 통합 아키텍처 기반의 동시 설계 — 고비용 칩도 로보택시 수익으로 상쇄하는 전략적 구조 <Tesla AI5 설계 요약> •제조 공정 : TSMC 3nm N3P •연산 성능 : Elon Musk 언급 – AI4
8
42
134
With A19, instead of making the NPU larger for AI workloads, Apple added a neural accellerators in the GPU architecture. Given some AI workloads are better on the GPU vs. NPU this makes a lot of sense and gives more flexiblity to the Apple’s GPU for AI worklaods.
7
25
164
#elonmusk #DOGE #optimus #starlink #ai #smartphone #인포마켓 #강용운대표 영상 번역 ➡️ https://t.co/jFCOa9pgG8
https://t.co/W674v70eyI
0
0
0
NVIDIA CFO 코렛 크레스, 2025 골드만삭스 Communacopia + Tech 컨퍼런스 주요 발언 정리 https://t.co/zBnEjGS4dL
#NVDA #엔비디아 #테크컨퍼런스 #인포마켓 #강용운대표
0
0
0
Toast
🚨ELON MUSK: "I haven't been to DC since May. The government is basically unfixable. I applaud David (Sacks') noble efforts...but at the end of the day if you look at our national debt...if AI and robots don't solve our national debt, we're toast."
3K
6K
50K
Understand that strong performance by Nebius $NBIS and CoreWeave $CRWV equates to NVIDIA $NVDA stock going higher... NVIDIA holds a $3.96 billion stake in CoreWeave and approximately $110 million stake in Nebius, and both companies heavily utilize NVIDIA’s GPUs for their AI and
24
41
331
Our key takeaways from $NVDA CFO Colette Kress during the 2025 $GS Communacopia + Tech Conference: —Data center revenue (inclusive of compute and networking, excluding H20) grew 12% QoQ in Q2 FY26 and is projected to grow another 17% QoQ in Q3, even without China contributions:
16
22
101
@ronaldmannak probably something akin to tensor cores on Nvidia hardware. like a logic block on each gpu core that performs a "take a tile of matrix A, multiply by tile of B, add to accumulator tile C" operation https://t.co/bOkUocJCFy unlike the neural engine which requires a pre-compiled
forums.developer.nvidia.com
Hello, I wanted to explore the tensor cores and use them to multiply some matrices generated in a kernel. Now I came across this line from the docs: C++ warp matrix operations leverage Tensor Cores...
1
3
16
BREAKING: Elon Musk reveals SpaceX & Starlink aim to make mobile phones connect directly to Starlink satellites with new chipsets in ~2 years, enabling global high-bandwidth connectivity without regional carriers. “You’ll be able to watch videos anywhere on your phone”
1K
4K
28K
🚨 BREAKING: Microsoft is making its biggest move yet to reduce dependence on OpenAI Despite investing $13+ billion in OpenAI, they’re hedging their bets with Anthropic for Office 365 Here’s the inside story:
44
112
777
Reasons Why Hybrid Bonding Is Extremely Challenging for HBM Stacking https://t.co/pt4fAuBxIL
#HBM_stacking #HBM #hybridbonding #하이브리드본딩 #인포마켓 #강용운대표
0
0
0
Hitachi Energy will invest $457M in a new transformer facility in South Boston, Virginia, plus $500M+ in other investments to address critical grid equipment shortages. The new facility will produce large power transformers with 3-year lead times. Construction starts this year,
8
27
137
Some of today’s most advanced AI workloads, like code and video generation, demand context processing at unprecedented scale, often exceeding one million tokens. This is why NVIDIA is launching Rubin CPX: a GPU purpose‑built for the compute‑intensive context phase of inference.
19
38
190
Free high-speed Wi-Fi is now available on board #AirFrance flights. A stable, fast and secure experience by @Starlink, already deployed on several aircraft and gradually being rolled out across the entire fleet by the end of 2026. https://t.co/EbHq24Nuu4
208
655
3K