jimxjim @jim74255 - Twitter Profile

5 days ago

仕事をする時、特に遊びに出かけながら仕事をする時は、A4用紙の半分の大きささえあれば、HHKB Type-Sを持ち運ぶのに十分です YBSG @PFU_HHKB #HHKBあるあるキャンペーン #HHKB30周年 #HHKB #HHKB30thanniversary

jim74255's tweet photo. 仕事をする時、特に遊びに出かけながら仕事をする時は、A4用紙の半分の大きささえあれば、HHKB Type-Sを持ち運ぶのに十分です
YBSG @PFU_HHKB

#HHKBあるあるキャンペーン
#HHKB30周年 #HHKB #HHKB30thanniversary https://t.co/X0XhQYDjLn

0

1

0

436

jimxjim

@jim74255

5 days ago

@ry__ku_14 会社の席にいる時も一人で川辺を散歩している時も忙しくて深夜まで働いたのに眠れない時もいつも「ブレス」を聴いています

1

0

50

jimxjim

@jim74255

6 days ago

@jimmy_yoasobi 吉米，這次還是一樣嗎

0

152

jim74255 retweeted

法克魷吉米 Jimmy | Real TSMC Insight | 純愛を学び中 💜

@jimmy_yoasobi

10 days ago

Long $SNDK $MU >>>June Market Chaos Major Correction in Progress – Buy in Batches Memory Super Cycle Far From Over AI Demand Will Continue to Explode Five Core Conclusions >Higher efficiency will not reduce memory demand Instead it will massively expand the overall market pie >Historical patterns repeatedly prove Jevons Paradox still holds true in the AI era >China's supply impact mainly affects mid-to-low-end segments HBM and advanced DRAM will see almost no effect in the short term >Once enterprises adopt AI they cannot stop Usage depth and breadth keep expanding >The 2028 "peak end" view is overly optimistic Real capital expenditure cycle will far exceed this timeline >>Higher efficiency Lowers AI usage cost -- Lower cost Encourages more industries and more companies to adopt at scale More adoption Requires even more advanced memory This is a powerful positive loop Not the bearish fantasy of "if it gets expensive people will stop buying" -- Look back at history In the 1990s when internet speed increased People didn't use the internet less They started watching videos playing games and data volume exploded In the 2000s when mobile data prices dropped People didn't make fewer calls They started mobile browsing watching videos and video calls After cloud computing costs fell Enterprises didn't buy fewer servers They moved SaaS big data and machine learning onto the cloud Every time efficiency improves The entire market scale grows exponentially The AI era is exactly the same -- From HBM3 to HBM4 Bandwidth increases dramatically The same GPU can process more tokens Token price per unit drops This attracts hesitant companies to enter in force Longer context more complex models more users Demand doesn't decrease It explodes across the board -- China's supply is indeed increasing But CXMT mainly produces DDR4 DDR5 standard memory YMTC focuses on NAND What AI truly needs is high-end HBM advanced DDR5 LPDDR5X China still faces huge gaps in technology equipment yield and packaging in these areas It will be difficult to catch up in the next 2-3 years or even longer -- The conclusion is simple Bears' biggest misjudgment Is treating "efficiency improvement" as "demand reduction" History has never supported this conclusion And the AI era will be no exception Memory GPU and power demand Will all grow exponentially together The super cycle will continue

jimmy_yoasobi's tweet photo. Long $SNDK $MU
>>>June Market Chaos
Major Correction in Progress – Buy in Batches

Memory Super Cycle Far From Over AI Demand Will Continue to Explode

Five Core Conclusions

>Higher efficiency will not reduce memory demand Instead it will massively expand the overall market pie

>Historical patterns repeatedly prove Jevons Paradox still holds true in the AI era

>China's supply impact mainly affects mid-to-low-end segments HBM and advanced DRAM will see almost no effect in the short term

>Once enterprises adopt AI they cannot stop Usage depth and breadth keep expanding

>The 2028 "peak end" view is overly optimistic
Real capital expenditure cycle will far exceed this timeline

>>Higher efficiency Lowers AI usage cost

--

Lower cost Encourages more industries and more companies to adopt at scale

More adoption Requires even more advanced memory

This is a powerful positive loop

Not the bearish fantasy of "if it gets expensive people will stop buying"

--

Look back at history

In the 1990s when internet speed increased People didn't use the internet less They started watching videos playing games and data volume exploded

In the 2000s when mobile data prices dropped People didn't make fewer calls They started mobile browsing watching videos and video calls

After cloud computing costs fell Enterprises didn't buy fewer servers They moved SaaS big data and machine learning onto the cloud

Every time efficiency improves The entire market scale grows exponentially

The AI era is exactly the same

--

From HBM3 to HBM4 Bandwidth increases dramatically

The same GPU can process more tokens Token price per unit drops

This attracts hesitant companies to enter in force Longer context more complex models more users

Demand doesn't decrease It explodes across the board

--

China's supply is indeed increasing

But CXMT mainly produces DDR4 DDR5 standard memory YMTC focuses on NAND

What AI truly needs is high-end HBM advanced DDR5 LPDDR5X

China still faces huge gaps in technology equipment yield and packaging in these areas It will be difficult to catch up in the next 2-3 years or even longer

--

The conclusion is simple

Bears' biggest misjudgment Is treating "efficiency improvement" as "demand reduction"

History has never supported this conclusion And the AI era will be no exception

Memory GPU and power demand Will all grow exponentially together

The super cycle will continue

2

20

7

9

38K

Who to follow

Graphtronauts

@graphtronauts_c

Home of long-term $GRT holders in @graphprotocol ecosystem | We run an indexer at @graphtronauts_x and build tools for the ecosystem → https://t.co/CZJaGRX9mD

Nick Hansen

@TwentyTwoNode

Team Lead at The Graph Foundation

Ethereum Daily

@EthereumDaily_

Your Daily Briefing on Ethereum News.

jim74255 retweeted

緑黄色社会 @ryokushaka

about 2 months ago

＼🏮緑黄色大夜祭🏮／ 2026.4.25 #緑黄色大夜祭2026 DAY1 出演者のみなさまと📸 最高の時間をありがとうございました🏮✨ Photo by @AZUSA_TAKADA 　　　　 @西槇太一 #緑黄色社会 #リョクシャカ

15

2K

256

60

76K

jimxjim

@jim74255

about 2 months ago

最高の祭り！ #緑黄色大夜祭 #リョクシャカ #緑黄色社会

0

3

0

1

588

jimxjim

@jim74255

2 months ago

#緑黄色夜祭 #リョクシャカ #緑黄色社会 #ZeppOsakaBayside

0

4

0

278

jimxjim

@jim74255

3 months ago

livestone vol.4 ついにLivestoneに参加できたこれまでの抽選では何度か外れていた🫠 番号がこれほど前になるのは初めてだ MCにはおそらく60～70％程度しか理解できていないと思いますがそれでもlivestoneの特別な魅力を感じることができます次回、大阪に緑黄色夜祭！ #リョクシャカ #livestone #緑黄色社会