Jukan
RT @eliant_capital: @GordianKnotDev There’s been rumors of around 150-170B in funding for OpenAI in the next raise. They’ll be fully funded through 2030 if that’s confirmed and entire complex will rip
tweet
RT @eliant_capital: @GordianKnotDev There’s been rumors of around 150-170B in funding for OpenAI in the next raise. They’ll be fully funded through 2030 if that’s confirmed and entire complex will rip
tweet
Jukan
Samsung Galaxy S27 Ultra Expected to Feature Polar ID Facial Recognition Technology
• Reports indicate that Samsung plans to implement the next-generation facial recognition technology called Polar ID in the Galaxy S27 Ultra. Unlike the existing Face ID system that relies on 3D depth modeling, Polar ID uses polarized light to recognize faces. This eliminates the need for large sensor space on the display, making it possible to achieve a true full-screen design without notches, pill-shaped cutouts, or large hole punches (citing Chinese media)
tweet
Samsung Galaxy S27 Ultra Expected to Feature Polar ID Facial Recognition Technology
• Reports indicate that Samsung plans to implement the next-generation facial recognition technology called Polar ID in the Galaxy S27 Ultra. Unlike the existing Face ID system that relies on 3D depth modeling, Polar ID uses polarized light to recognize faces. This eliminates the need for large sensor space on the display, making it possible to achieve a true full-screen design without notches, pill-shaped cutouts, or large hole punches (citing Chinese media)
tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
Learnt something new today: "Bloomberg's default index P/E calculation includes the negative earnings of loss-making companies in the denominator. This reduces the total "Earnings" pool, causing the P/E multiple to spike." https://t.co/OI3Rmhu0kf
tweet
Learnt something new today: "Bloomberg's default index P/E calculation includes the negative earnings of loss-making companies in the denominator. This reduces the total "Earnings" pool, causing the P/E multiple to spike." https://t.co/OI3Rmhu0kf
tweet
Offshore
Photo
God of Prompt
RT @godofprompt: vertex AI just leaked "fennec" in an error log
claude sonnet 5 reportedly drops tomorrow
the rumored specs:
> 50% cheaper than opus 4.5
> still 1M context window but faster
> can spawn parallel sub-agents from terminal
> allegedly hitting 80.9% on SWE-bench
the wildest part: "dev team mode" where you give a brief and agents build the full feature autonomously
treat this as unverified.
but if real, coding agents just changed overnight.
tweet
RT @godofprompt: vertex AI just leaked "fennec" in an error log
claude sonnet 5 reportedly drops tomorrow
the rumored specs:
> 50% cheaper than opus 4.5
> still 1M context window but faster
> can spawn parallel sub-agents from terminal
> allegedly hitting 80.9% on SWE-bench
the wildest part: "dev team mode" where you give a brief and agents build the full feature autonomously
treat this as unverified.
but if real, coding agents just changed overnight.
tweet
Offshore
Video
God of Prompt
This was 6 years ago.
Wondering how we’ll look back on clawdbot 6 years from now on.
tweet
This was 6 years ago.
Wondering how we’ll look back on clawdbot 6 years from now on.
I just built a *functioning* React app by describing what I wanted to GPT-3.
I'm still in awe. https://t.co/UUKSYz2NJO tweet
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
1/ Is the "Korea Discount" finally coming to an end?
South Korea has been a value trap for decades, due to a low return on equity. But reforms are finally starting to move the needle. https://t.co/D3ajVt4z9O
tweet
1/ Is the "Korea Discount" finally coming to an end?
South Korea has been a value trap for decades, due to a low return on equity. But reforms are finally starting to move the needle. https://t.co/D3ajVt4z9O
tweet
Offshore
Photo
God of Prompt
Claude might be getting image model?
tweet
Claude might be getting image model?
@AndrewCurran_ @repligate @anthrupad 👀 https://t.co/UdDPN5yrCV tweet
Michael Fritzell (Asian Century Stocks)
AI tools might be great at generating text. But some of the recent deep dives on Substack are less than impressive. Even if you go this route, you probably need to tell ChatGPT to match your way of writing...
Or do what I do: write yourself, and just use AI for inspiration.
tweet
AI tools might be great at generating text. But some of the recent deep dives on Substack are less than impressive. Even if you go this route, you probably need to tell ChatGPT to match your way of writing...
Or do what I do: write yourself, and just use AI for inspiration.
tweet
The Transcript
$NVDA
tweet
$NVDA
We love working with NVIDIA and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time.
I don't get where all this insanity is coming from. tweet
X (formerly Twitter)
Sam Altman (@sama) on X
We love working with NVIDIA and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time.
I don't get where all this insanity is coming from.
I don't get where all this insanity is coming from.
Offshore
Photo
Michael Fritzell (Asian Century Stocks)
Now I know why insiders are buying shares in Ultrajaya
cc: @aktieblogg
https://t.co/BILf3b4ERo
tweet
Now I know why insiders are buying shares in Ultrajaya
cc: @aktieblogg
https://t.co/BILf3b4ERo
tweet
Offshore
Photo
Jukan
Samsung Leads in HBM4 Performance, but Nvidia Backs SK Hynix for Its 'Yield'
Samsung Electronics has launched a counteroffensive in the 6th-generation High Bandwidth Memory (HBM4) market by achieving the highest performance among the top three memory makers. However, SK Hynix is expected to maintain its market dominance, driven by its superior production yields.
As global big tech companies like Nvidia face extreme HBM supply shortages, they are prioritizing the securing of stable volumes over raw performance. In this environment, "yield competitiveness" has emerged as the decisive factor in the competition.
■ Performance (Samsung) vs. Yield (Hynix): Nvidia’s Choice
According to semiconductor industry sources on the 3rd, Samsung Electronics, which begins mass production of HBM4 this month, is recognized for its edge in performance.
By simultaneously applying the 1c process (10nm-class 6th generation) and its 4nm foundry process, Samsung has pushed HBM4 speeds to 11.7Gbps. This surpasses the 10–11Gbps range achieved by its competitor, SK Hynix, making it the highest performance in the industry.
In contrast, SK Hynix experienced delays in passing quality (qualification) tests after undergoing a redesign process for some products late last year to meet Nvidia's upgraded performance requirements. It is understood that SK Hynix currently trails Samsung in terms of absolute speed.
Despite this, industry analysts predict that SK Hynix will secure 50–60% or more of Nvidia’s initial HBM4 orders. The secret behind SK Hynix maintaining its "first vendor" status despite the performance gap lies in its overwhelming yield rates. SK Hynix already achieved yields of 80–90% with HBM3E and aims to secure similar levels for HBM4 by utilizing its proprietary packaging technology.
■ "Negligible Speed Difference... Stable Supply is Top Priority"
The reason big tech firms are fixated on yield is the unprecedented HBM supply shortage.
Not only Nvidia but also AMD, Google, and Microsoft are launching new chips that require massive amounts of HBM, causing demand to outstrip supply for both HBM4 and HBM3E.
Nvidia CEO Jensen Huang recently highlighted this imbalance, stating, "AI requires memory, and we need a significant amount of memory semiconductors this year. Because demand is much higher this year, the memory supply chain is in a difficult situation." Both Samsung and SK Hynix also officially confirmed during their Q4 earnings calls that HBM supply is failing to keep up with demand.
An industry insider familiar with the matter explained, "For customers, the performance difference between 10.5Gbps and 11Gbps is not significantly felt in real-world environments. Rather, if performance is comparable, the ability to receive a stable supply of promised volumes at the designated time is a much more critical value."
In this situation, Samsung Electronics' yield for the 1c DRAM used in HBM4 is reportedly still stuck around 60%, and the final yield after packaging is estimated to be even lower. A source familiar with Samsung's internal situation noted, "The yield for HBM-grade 1c has not yet surpassed 70%; I understand it is currently between 50% and 60%." While 1c DRAM is split between general-purpose and HBM use, the HBM-grade 1c has lower yields due to higher manufacturing complexity.
Low yields not only increase production costs but also make it difficult to meet large-scale orders on time. Furthermore, since Nvidia requires a variety of HBM speeds (ranging from 10Gbps to over 11Gbps) rather than just the fastest units, SK Hynix is expected to continue occupying more than half of the market.
■ Intensifying Market Share Battle... Second Round Expected in HBM4E
Current market forecasts suggest SK Hynix will take over half of the HBM4 market, while Samsung Electronics is expected to capture 30–40% of the volume. However, experts suggest that if Samsung succeeds in stabilizing its HBM4 yield to over 80% within the year, it could rapidly increase its m[...]
Samsung Leads in HBM4 Performance, but Nvidia Backs SK Hynix for Its 'Yield'
Samsung Electronics has launched a counteroffensive in the 6th-generation High Bandwidth Memory (HBM4) market by achieving the highest performance among the top three memory makers. However, SK Hynix is expected to maintain its market dominance, driven by its superior production yields.
As global big tech companies like Nvidia face extreme HBM supply shortages, they are prioritizing the securing of stable volumes over raw performance. In this environment, "yield competitiveness" has emerged as the decisive factor in the competition.
■ Performance (Samsung) vs. Yield (Hynix): Nvidia’s Choice
According to semiconductor industry sources on the 3rd, Samsung Electronics, which begins mass production of HBM4 this month, is recognized for its edge in performance.
By simultaneously applying the 1c process (10nm-class 6th generation) and its 4nm foundry process, Samsung has pushed HBM4 speeds to 11.7Gbps. This surpasses the 10–11Gbps range achieved by its competitor, SK Hynix, making it the highest performance in the industry.
In contrast, SK Hynix experienced delays in passing quality (qualification) tests after undergoing a redesign process for some products late last year to meet Nvidia's upgraded performance requirements. It is understood that SK Hynix currently trails Samsung in terms of absolute speed.
Despite this, industry analysts predict that SK Hynix will secure 50–60% or more of Nvidia’s initial HBM4 orders. The secret behind SK Hynix maintaining its "first vendor" status despite the performance gap lies in its overwhelming yield rates. SK Hynix already achieved yields of 80–90% with HBM3E and aims to secure similar levels for HBM4 by utilizing its proprietary packaging technology.
■ "Negligible Speed Difference... Stable Supply is Top Priority"
The reason big tech firms are fixated on yield is the unprecedented HBM supply shortage.
Not only Nvidia but also AMD, Google, and Microsoft are launching new chips that require massive amounts of HBM, causing demand to outstrip supply for both HBM4 and HBM3E.
Nvidia CEO Jensen Huang recently highlighted this imbalance, stating, "AI requires memory, and we need a significant amount of memory semiconductors this year. Because demand is much higher this year, the memory supply chain is in a difficult situation." Both Samsung and SK Hynix also officially confirmed during their Q4 earnings calls that HBM supply is failing to keep up with demand.
An industry insider familiar with the matter explained, "For customers, the performance difference between 10.5Gbps and 11Gbps is not significantly felt in real-world environments. Rather, if performance is comparable, the ability to receive a stable supply of promised volumes at the designated time is a much more critical value."
In this situation, Samsung Electronics' yield for the 1c DRAM used in HBM4 is reportedly still stuck around 60%, and the final yield after packaging is estimated to be even lower. A source familiar with Samsung's internal situation noted, "The yield for HBM-grade 1c has not yet surpassed 70%; I understand it is currently between 50% and 60%." While 1c DRAM is split between general-purpose and HBM use, the HBM-grade 1c has lower yields due to higher manufacturing complexity.
Low yields not only increase production costs but also make it difficult to meet large-scale orders on time. Furthermore, since Nvidia requires a variety of HBM speeds (ranging from 10Gbps to over 11Gbps) rather than just the fastest units, SK Hynix is expected to continue occupying more than half of the market.
■ Intensifying Market Share Battle... Second Round Expected in HBM4E
Current market forecasts suggest SK Hynix will take over half of the HBM4 market, while Samsung Electronics is expected to capture 30–40% of the volume. However, experts suggest that if Samsung succeeds in stabilizing its HBM4 yield to over 80% within the year, it could rapidly increase its m[...]