Discussion Should MLID be considered a legitimate source

[Saylick]

MLID saying ARL 8+32 is coming is a major red flag. That SKU is deader than dead, as far as I understand.

 

[igor_kavinski]

MLID saying ARL 8+32 is coming is a major red flag.

I think Intel can only do 32 Skymonts in the same area if they placed a 2nd 16-core Skymont die on top of the existing 16-core die. But then there might be a clock regression to control all that heat build up.

 

[marees]

Looks like, no.

No mainstream new site carried his leak of sneak RDNA 4 launch in Q4 this year

 

[511]

@Hitman928 It doesn't have arc is cancelled in the screenshot it shouldn't be marked as a spoiler as well.

 

[Hitman928]

@Hitman928 It doesn't have arc is cancelled in the screenshot it shouldn't be marked as a spoiler as well.

I’m not sure what you are saying. Are you saying I should include a screenshot of his Arc is canceled mis-prediction?

 

[511]

I’m not sure what you are saying. Are you saying I should include a screenshot of his Arc is canceled mis-prediction?

YES

 

[Hitman928]

YES

Gotcha. When I did the analysis, I took what was purported to be some of his better slides/predictions to review to be as fair to him as possible. There are numerous examples that ended up much worse than what I showed, some of which he’s deleted after reality hit. Others in the thread added some examples and anyone is free to add more but the initial post already took way longer than I wanted to give to the effort so I wasn’t going to keep adding to the original post.

 

[marees]

I tried to watch this video 3 times today

All the 3 times, I fell asleep

we need AI advancement to summarize pointless videos

https://twitter.com/x/status/1956125577271730369

 

[DAPUNISHER]

I tried to watch this video 3 times today

All the 3 times, I fell asleep

we need AI advancement to summarize pointless videos

https://twitter.com/x/status/1956125577271730369

Tweak Town has got you homie - https://www.tweaktown.com/news/1070...p-to-10-faster-than-285k-in-gaming/index.html

I should have mentioned if you want to know about the Zen 7 stuff, read the comments. It'll stay on AM5 is the claim.

Honestly I don't bother with anything from Tom but the shows with interesting guests.

 

[marees]

Tweak Town has got you homie - https://www.tweaktown.com/news/1070...p-to-10-faster-than-285k-in-gaming/index.html

I should have mentioned if you want to know about the Zen 7 stuff, read the comments. It'll stay on AM5 is the claim.

Honestly I don't bother with anything from Tom but the shows with interesting guests

I was trying to see what he had to say on xbox next & rdna 5 but only thing I got was that he expects a refresh of rdna 4

Not going to watch it a 4th time.

 

[DAPUNISHER]

I doubt you missed anything. Usually another outlet will do what TT did and summarize anything clickbaity.

 

[igor_kavinski]

Not going to watch it a 4th time.

That's why I always copy the transcript and give it to ChatGPT.

People who can tolerate watching him for more than a minute are not normal.

 

[Ranulf]

That's why I always copy the transcript and give it to ChatGPT.

People who can tolerate watching him for more than a minute are not normal.

Im usually gaming with it playing the background.

 

[Hulk]

I find it very, very telling that MLID has recently started showing old slide for like 1 second, picking out 1 "leak" that was correct and staying, "See, this is what I leaked 3 years ago?"

Meanwhile the entire rest of the slide is completely wrong. You know what they say about a broken clock.

Anyway, my point is that he knows people around the net are laughing at his ridiculous "leaks" because 90% of them end up being complete BS so he's cherry picking some "easy hits" to validate himself. It's a little sad and desperate actually.

He needs to remember he is not "real" wrestling, he is WWE, "Entertainment," not reality, and he is good at the entertainment business. Stay in your lane and you'll be fine. Try and convince us this is real wrestling and the gig is up.

 

[Doug S]

That's why I always copy the transcript and give it to ChatGPT.

People who can tolerate watching him for more than a minute are not normal.

Come to think of it, one good use for AI would be if you could tell it to watch a particular youtube video and give you a summary of what it said, including visual aids referenced by the summary. That way you avoid all the stupid sponsor plugs, advertising (if you actually see ads on youtube, I still don't) and other wasted time. That's the main reason I almost always ignore any link someone references that points at youtube. With text I can skim and get to the stuff I care about it. That's a lot more of a pain with a video, even when watching it with subtitles at 2.5x.

I have to imagine Google would be rather upset if this became a popular way of consuming youtube content lol!

 

[fastandfurious6]

make 200 claims, 20 are correct, "see what i predicted 10 years ago!!"

 

[igor_kavinski]

The really incredulous stuff is highlighted and bolded in RED.

Arrow Refresh

• Not a Zen 6 competitor; not a full Alder Lake–to–Raptor Lake type jump.
• Clocks: higher core clocks announced, but not dramatic (expect mid-6 GHz range, not 7 GHz).
• Memory/IO: ring bus clock increases are a focus; die-to-die clocks are being raised to ease bottlenecks.
• Overall performance uplift: about 7–10% in gaming; roughly 5–10% across many apps, with potential edge cases above 10%.
• Power: TDP is not expected to rise much; no major jump in power targets.
• Market stance: stronger gaming performance than Arrow Lake 1.0, but not a Zen 6 killer. Marketing push planned to position it against Zen 5.

Zen 5 Refresh

• RAM/memory: better RAM support planned, driven by binning the memory controllers for the new chips; faster standard RAM expected.
• Clocks/RAM combo: higher clocks likely; faster RAM pairing may enable stronger performance.
• Dual VCache possibility: a dual-VCache variant (e.g., a 9950X60-type SKU) has been discussed but not independently verified.
• Launch window: expected by Q1 next year; positioned as a strong response to Arrow Refresh and a step up in gaming/IPC.
• Overall: not a Zen 6 upgrade, but a meaningful refresh that should keep AMD competitive, especially in gaming; a potential bump to bring Zen 5 closer to Zen 5 X3D in some areas, though it won’t instantly overtake it.

I don't see the stuff in red happening on N3B. Has to be a process change.

 

[511]

6Ghz+ is crap the TDP one might make sense if they improv DLVR

 

[coercitiv]

I don't see the stuff in red happening on N3B. Has to be a process change.

You know how this works: he can't just say 5.9-6Ghz boost, he needs his audience engaged. So first he says mid-6 Ghz and everybody raises eyebrows, rubs palms. THEN, just before the refresh launch, he comes back with a second leak: Intel changed boost target, now it's low-6Ghz due to "insert reason" >> they'll fix it with Nova Lake!

The process is matured, yields are likely better, binning is more efficient as a result. Intel likely tuned on their side as well. For me the higher boost clocks are not as important as the higher uncore clocks (assuming MLID isn't completely full of it), that should help with performance consistency. Might help peak SSD performance too.