You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
I am just saying he knows things and details about supply chain and what people are doing. I assume he has a great network as some of the details he knows are incredibly accurate and not widely known. As Tyrion Lannister said "That's what I do. I drink and I know things" (I dont know if Dylan drinks LOL)
I dont know his education or background. I am not hiring him to lead R&D. I just listen and then double check.
There are several people posting on their websites who aren't deeply expert in the fields they're posting about, but because some of them are very well-connected in the industry they often give useful pointers to what's going on. For example, in Dylan's case he wrote an article about Directed Self-Assembly with High-NA EUV lithography. I asked around with a couple of people in Intel, one of them chuckled when I asked about the article, but DSA is apparently real, but more difficult to implement than Dylan's article makes it out to be. The value was I saw it first from him. Dylan has also started using co-authors lately, and I've noticed a modest improvement in technical quality, though to be honest I'm not expert enough to really judge them on my own.
I read The Next Platform, mostly written by Timothy Morgan, and I do have enough expertise to judge his writing (and my judgement is not positive, except when he's doing financial analyses), but, again, he is very well connected and points me towards news items to further investigate without his, uh, technical analysis. Another analyst who pops to mind is Patrick Moorhead, but he stays at a high level, targets non-technical people, and is very useful to get a feel for what's going on. He's also very polished and keeps to what he knows, which is admirable.
On Friday April 12th Intel held a press briefing on their adoption of High NA EUV with Intel fellow and director of lithography Mark Phillips. In 1976 Intel built Fab 4 in Oregon, the first Intel fab outside of California. With the introduction of 300mm Oregon became the only development site...
There are several people posting on their websites who aren't deeply expert in the fields they're posting about, but because some of them are very well-connected in the industry they often give useful pointers to what's going on.
Sounds allot like Tom over at MLID. I think he worked in automotive supply chain management prior to becoming a full time youtuber and just got to know allot of people in the cpu / gpu supply chain.
Dylan is clearly well connected to the supply chain and gets a lot of non-public data. Technical depth is another matter.
To Dylan's credit I have to say he is not one of the "Industry analysts" who are basically marketing extensions of companies and add zero value. Unfortunately, there are far too many industry analysts and influencers lately as that is the easy money path.
I am almost 100% certain he has no contacts and no sources. The only correct thing I have ever seen fall out of his mouth is what you could have gotten from Wikipedia, a college textbook, and or a basic understanding of science/engineering. The data from his so called “team” or “sources” is clearly made up. Darts against the dartboard if you will. His stuff ranges from proven false by public data/statements, or would require multiple different people at different companies/areas disclosing trade secrets (a network that I don’t think any one person, least of all Dylan, has). An example of the later was Dylan talking about Kokosai thermal ALD tools as being some specific amount higher throughput than ALL single wafer tools (albeit I don’t remember the exact % he claimed). A claim like that would need engineers who work on every ALD tool and on every application for such a claim to be made. And I don’t think I need to tell you that this methodology wouldn’t produce one easy number. If it did somebody would have 100% MSS. It also is a useless claim because different thermal tools operate in different pressure and temperature regimes. Then there are plasma enhanced ALD tools. TLDR batch vs single wafer is one of many factors impacting WPH and all single wafer tools won’t perform the same as they don’t all work the same. More total cycles also might favor different tools than short total cycle tools. Dirtier processes might push you away from tools that are more susceptible or that have long chamber cleans. The fact I can catch stuff like that without ever working on an ALD tool or even doing a tool selection before in my opinion throws alot of red flags.
Technical depth is another matter.
To Dylan's credit I have to say he is not one of the "Industry analysts" who are basically marketing extensions of companies and add zero value. Unfortunately, there are far too many industry analysts and influencers lately as that is the easy money path.
I would say he is valueless. He makes some crazy logical leaps that don’t even match his own “analysis” or actual real numbers. It also really grinds my gears seeing his watermark get plastered on other folks materials. Even in instances where he does reference the original source (and not just himself or his “team”) it feels one or two steps removed from plagiarism to me.
It also doesn’t help that feigning knowledge where there is only hot air is also not a great way to earn my respect. Someone like Asianometry does cite all of his sources and references (even graphics) and will admit when he doesn’t understand something very well or at all. In my opinion John is what these sort of folks should strive for rather than just vomiting out self referential garbage to mislead others for a quick buck.
I am almost 100% certain he has no contacts and no sources. The only correct thing I have ever seen fall out of his mouth is what you could have gotten from Wikipedia, a college textbook, and or a basic understanding of science/engineering. The data from his so called “team” or “sources” is clearly made up. Darts against the dartboard if you will. His stuff ranges from proven false by public data/statements, or would require multiple different people at different companies/areas disclosing trade secrets (a network that I don’t think any one person, least of all Dylan, has). An example of the later was Dylan talking about Kokosai thermal ALD tools as being some specific amount higher throughput than ALL single wafer tools (albeit I don’t remember the exact % he claimed). A claim like that would need engineers who work on every ALD tool and on every application for such a claim to be made. And I don’t think I need to tell you that this methodology wouldn’t produce one easy number. If it did somebody would have 100% MSS. It also is a useless claim because different thermal tools operate in different pressure and temperature regimes. Then there are plasma enhanced ALD tools. TLDR batch vs single wafer is one of many factors impacting WPH and all single wafer tools won’t perform the same as they don’t all work the same. More total cycles also might favor different tools than short total cycle tools. Dirtier processes might push you away from tools that are more susceptible or that have long chamber cleans. The fact I can catch stuff like that without ever working on an ALD tool or even doing a tool selection before in my opinion throws alot of red flags.
I would say he is valueless. He makes some crazy logical leaps that don’t even match his own “analysis” or actual real numbers. It also really grinds my gears seeing his watermark get plastered on other folks materials. Even in instances where he does reference the original source (and not just himself or his “team”) it feels one or two steps removed from plagiarism to me.
It also doesn’t help that feigning knowledge where there is only hot air is also not a great way to earn my respect. Someone like Asianometry does cite all of his sources and references (even graphics) and will admit when he doesn’t understand something very well or at all. In my opinion John is what these sort of folks should strive for rather than just vomiting out self referential garbage to mislead others for a quick buck.
I cannot speak to your specific example and I have no way of evaluating it but what I noticed, especially on the more recent supply chain ramps on GPUs is that he gets good data and clearly what he is doing is not possible without contacts without knowledge - at hypers and in Taiwan supply chain. As far as his analysis goes, I disagree a bit. It is definitely not valueless. He makes a few good points though buried in a lot of technical mumbo jumbo which only non-technical people, like the finance types who are likely his main customers, would like. This technical info is also clearly from sources (again, well connected) but the lack of depth shows as his connecting of the dots is weak. It is a matter of time though. If he keeps at it, another few more years I think he will get quite good.
