To keep this topic from polluting other threads, I've created its own.
@dullard
Here is your list that you wanted to discuss:
Nov 2020
June 2021
June 2022
Nov 2022
Dec 2022
May 2023
June 2023
Now this took me way to long to go through but this is a big reason why myself and most everybody else does not see him as legitimate, in addition to just hearing him talk on subjects and quickly realizing he has no clue what he's actually talking about. Feel free to discuss but I'm out for a bit as I now have real work to catch up on.
@dullard
Here is your list that you wanted to discuss:
Nov 2020
June 2021
June 2022
Nov 2022
Dec 2022
May 2023
June 2023
- I listed this as wrong because MTL not being Ocean Cove was known well before this "leak" and he continued to report it as Ocean Cove well after others knew it wasn't. So he is just trying to save face here by claiming he basically knew it was RWC all along but just called it Ocean Cove because his source wanted him to.
- I don't think there's any real way to confirm that RWC was actually designed to be node agnostic so I'm not listing it as wrong, but others say that it isn't. Even if it was, this is not a leak. Intel themselves put out press slides the year before this "leak" talking about how future design work would be node agnostic and that designs would have opportunities to be fabricated on multiple nodes if needed. Others have reported that this didn't really start until ARL.
- MTL was not 3D stacked. You may give him credit for saying it would utilize an IO die (possibly from TSMC) but if you look at this "leak" he predicts that it could be 3D stacked, and/or utilize a separate IO die and that die might be from TSMC, or it could be monolithic. He literally puts out every possible configuration it could be so no matter what, he could claim he was right. He does not get credit for any part of this.
- ADL was completely monolithic. You're probably giving him credit for "leaking" that MTL was a tiled design but once again, this was known before his leak. In fact, it was tweeted out by an Intel executive before MLID even made his video.
- Crestmont's name had been bouncing around for months before MLID's video on it.
- Maybe this is the best prediction he gets most credit for but Intel had been talking about adding AI type blocks into their SoCs before he leaked it.
- Again, known before his video. We even had die shots of MTL by this point.
- See here.
- MTL does not beat RPL across the board, not even close. It is more efficient but, well, duh.
- Finally concedes one month after his previous update video that MTL desktop doesn't exist but still hints that it just might come out in 2024. Only thing correct here is that 6p+8e is the top die but that had already been known for a long time by this point.
- Already known
- Already leaked
- Already known for at least a year.
Now this took me way to long to go through but this is a big reason why myself and most everybody else does not see him as legitimate, in addition to just hearing him talk on subjects and quickly realizing he has no clue what he's actually talking about. Feel free to discuss but I'm out for a bit as I now have real work to catch up on.
Last edited: