SemiAnalysis ✓
@SemiAnalysis_ · joined 10 Jan 2024
165 863Followers
34Following
3 331Posts total
3.8MViews on collected posts
Son gönderiler
Yeni gönderiler aranıyor…
@SemiAnalysis_ You invented the name, not the industry.
Take a W with some class.
645 views · 15 likes
· 0 reposts · 3 replies
Open on X →
Neocloud is now an entire industry. The word was coined by a SemiAnalysis researcher. Some of the companies using it still have no idea.
"Nobody believes me when I say that. I'll tell people externally, yeah, we're the ones who invented the word, and they're like, No, you https:
56.6K views · 204 likes
· 13 reposts · 31 replies
Open on X →
1:06
@SemiAnalysis_ Is this a market top signal?
5.1K views · 95 likes
· 0 reposts · 0 replies
Open on X →
@SemiAnalysis_ What is wrong with you guys?
10.4K views · 169 likes
· 0 reposts · 6 replies
Open on X →
@SemiAnalysis_ @grok is this bullish or bearish or neither for mu nvda
2.7K views · 6 likes
· 0 reposts · 2 replies
Open on X →
GPT-6 Astra reportedly gets deeper without getting bigger. The labs already know what that means for scaling.
"It's basically confirmed that GPT-6 Astra uses loop transformers, which means that instead of adding parameter count, it goes through the layers more than once. You add
121.3K views · 1.1K likes
· 72 reposts · 57 replies
Open on X →
0:41
@SemiAnalysis_ this is a silly argument. GPT-5.6 Sol was 88.8% on TerminalBench 2.1 and 37.3% on TerminalBench 4.0. That doesn't make it a more benchmaxxed model than GPT-6 Astra even though the diff is larger. TerminalBench 4.0 is simply a much harder eval that isn't saturated,
90K views · 1.5K likes
· 45 reposts · 61 replies
Open on X →
The solution is more high quality private benchmarks. If you have a great private benchmark, reach out to @maxkan. We'd love to chat! (5/5)
22.3K views · 195 likes
· 1 reposts · 6 replies
Open on X →
Ultimately, this is the fate of all good public benchmarks. TB 4.0 is no exception. It’s only useful signal now because it was released 2 weeks ago. Since all the tasks are similarly public, it won't be long until it's hillclimbed by all the aspiring “frontier” labs. (3/5)
77.1K views · 239 likes
· 4 reposts · 1 replies
Open on X →
How is this possible? All of the tasks in Terminal Bench 2.1 are fully public. Though Meta and Google would never train on the tasks directly, they absolutely will buy data from RL env startups that’s designed to mimic TB 2.1 tasks as closely as possible. The net effect is the
27.6K views · 258 likes
· 3 reposts · 5 replies
Open on X →
Gemini 3.8 Flash on DeepSWE is another good example. Clearly, Datacurve made a ton of money selling Google DeepSWE-shaped tasks. (4/5) https://t.co/S6lf25GofC
26.1K views · 211 likes
· 2 reposts · 1 replies
Open on X →
Gemini 3.8 Flash and Muse Spark 1.3 are two of the most clearly benchmaxxed models we've seen yet. Despite being comparable to both GPT-6 and Fable 5.1 on Terminal Bench 2.1, their Terminal Bench 4.0 performance is markedly worse. (1/5)🧵 https://t.co/K2ccQjWm11
353K views · 2.3K likes
· 138 reposts · 96 replies
Open on X →
OPENAI<>NVIDIA: It is clearly a good idea for OpenAI to start an ASIC program, spend hundreds of millions on R&D, present at Hot Chips, deploy some stuff, scare NVIDIA, and get some more money and backstops from NVIDIA on the order of billions of dollars.
If Jalapeno does well h
26.1K views · 145 likes
· 6 reposts · 15 replies
Open on X →
Everyone is watching the federal/DOE-owned land at the center of the 8 GW Portsmouth site in the OpenAI–Nvidia deal. The March 2026 groundbreaking took place at the DOE’s Portsmouth Site, a former uranium-enrichment plant, and that’s where the attention has stayed. The federal ht
23.6K views · 127 likes
· 8 reposts · 3 replies
Open on X →
NVIDIA Research 🚀 has produced some great research, like LatentMoE (used in Kimi K3) and GatedDeltaNets (used in Qwen). But for e2e frontier training, NVIDIA's bureaucratic culture has produced embarrassing models like Nemotron3 Ultra.
Despite NVIDIA Research having amazing http
68.3K views · 307 likes
· 19 reposts · 34 replies
Open on X →
As the industry moves towards Co-Packaged and Near-Packaged Optics, the fiber attach is forced to land on the PIC itself. Fiber coupling is a very precise endeavor, with micron level precision requirements. An edge coupler wants the fiber core roughly within half a micron of the
40.7K views · 147 likes
· 8 reposts · 4 replies
Open on X →
@SemiAnalysis_ Attention to detail (regardless of how meaningless) is what makes SemiAnalysis so great. Can’t wait to drop Katsu is not curry on someone!
9.9K views · 22 likes
· 0 reposts · 1 replies
Open on X →
As educational as this post was, our recent deep dive on Jalapeno is even more educational. Read more here: (6/6) https://t.co/dbuAUrRSt1
23.6K views · 99 likes
· 2 reposts · 6 replies
Open on X →
Here is where it gets better. Katsu is short for katsuretsu, the Japanese transliteration of the English word "cutlet." A European dish adopted into Japan during the Meiji era, name included. So the word made a full round trip. Europe invents the cutlet. Japan borrows the word.
21.3K views · 59 likes
· 3 reposts · 2 replies
Open on X →
As the dish exploded in popularity in the 2010s, the name got compressed. "Chicken katsu curry" became "a katsu." The sauce became "katsu sauce." UK supermarkets now sell katsu-flavored everything, referring to the sauce and the cutlet nowhere in sight. (4/6) https://t.co/HiNuJwK
18.1K views · 42 likes
· 1 reposts · 1 replies
Open on X →
Katsu is a Japanese breaded fried cutlet. In Japan, katsu curry means exactly that: a katsu, served on curry. The word katsu refers only to the cutlet. So how did katsu come to mean "curry" to some people? That one traces back to the UK. Wagamama put chicken katsu curry on its ht
22K views · 61 likes
· 2 reposts · 2 replies
Open on X →
OpenAI went with a spicy theme for the codenames. The ASIC is a pepper codename, and various parts of the system are named after curries. Vindaloo for the ASIC trays, Chana for the switch trays, and Katsu for the host CPU trays. Except Katsu is NOT a curry. (2/6) https://t.co/iQt
29K views · 115 likes
· 8 reposts · 5 replies
Open on X →
There is a lot of deserved hype for how great OpenAI's Jalapeno ASIC and system will be. However, there is one major flaw that no one is talking about at all... (1/6)🧵 https://t.co/7mlrpDWAsp
191.5K views · 648 likes
· 29 reposts · 36 replies
Open on X →
@SemiAnalysis_ How the fuck did you manage to get this this early
1.2K views · 15 likes
· 0 reposts · 0 replies
Open on X →
@SemiAnalysis_ 1*C2-Ultra-4.24GHz
4*C2-Pro-3.74GHz
1*C1-Ultra-3.36GHz
4*C1-Pro-2.88GHz
Removed the wrong annotations. https://t.co/iNkYzxq6as
2.2K views · 21 likes
· 0 reposts · 0 replies
Open on X →
Aynı büyüklükteki hesaplara karşı
Son 90 güne ait 43 gönderi, 100K–1M takipçi aralığıyla yan yana. geniş kitleye gösteriliyor, ama izleyenlerin azı tepki veriyor.
Medyan görüntülenme22 293bu hesap5 983100K–1M için medyan
Erişim, %13.44%bu hesap1.49%100K–1M için medyan
Etkileşim, %0.45%bu hesap0.93%100K–1M için medyan
| Ölçüt | Bu hesap | 100K–1M için medyan | Oran |
|---|---|---|---|
| Gönderi başına medyan görüntülenme | 22 293 | 5 983 | 3.73× |
| Erişim (görüntülenme ÷ takipçi) | 13.44% | 1.49% | 9.02× |
| Etkileşim oranı | 0.45% | 0.93% | 0.49× |
Bu aralıktaki diğer hesaplar → Başka bir hesapla karşılaştır → Bu kıyas değerleri nasıl kuruluyor →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
27.6K8 Sep
77.1K
22.3K
90K
18.8K15 Sep
121.3K
2.7K
549.5K20 Sep
13.1K
10.4K
5.1K
56.6K21 Sep
16.4K
645
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.97%8 Sep
0.32%
0.91%
1.83%
0.09%15 Sep
1.00%
0.30%
0.19%20 Sep
0.32%
1.68%
1.88%
0.45%21 Sep
0.07%
2.79%
Reactions — likes, reposts, replies and quotes — divided by views. Median for 100K–1M accounts is 0.93%.
What the audience does
Likes75.6%9 005 in total
Reposts3.2%383 in total
Replies4.1%494 in total
Quotes1.7%200 in total
Bookmarks15.3%1 827 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Benzer hesaplar
ایران اینترنشنال ورزشی ✓
@iranintlsport
331.7K followers
· 48K posts
Everson Zoio👁 ✓
@EversonZoio
331.6K followers
· 12.7K posts
Radio 10 - AM 710
@Radio10
331.6K followers
· 349.7K posts
Osman Nuri Kabaktepe ✓
@osmannnurika
331.6K followers
· 12.3K posts
AZRA😳 ✓
@azra_lifts
331.6K followers
· 9.7K posts
Sonic the Hedgehog ✓
@SonicMovie
331.6K followers
· 19K posts
Steve · Millionaire Habits ✓
@SteveOnSpeed
331.6K followers
· 43.9K posts
President of the Senate
@SPNigeria
331.6K followers
· 7.5K posts