He managed to instill a culture of saying “no” to absolutely any risk taking. Anyone remotely competent to turn this company around in LTD is still paralyzed by this fear. Need to start over.
9 replies (most recent on top)
@1rxz: interesting point about the PTD managers bungling ramp. I've personally seen this mentality of showing "capability" rather than manufacturability. Does TSMC have ramp reporting to a separate org?
"the hot box craze cost us 7% production"
I would definitely assume that if people are hot boxing, their productivity will decrease and their time needed for snacking will increase. Not to mention the safety hazards.
If you know, you know. Polo
I was in Ramp from 2005 until it merged with PTD (engineering). I admired Sohail - he was in the building 18 hours a day and you’d see him in the Cafe at 10 at night talking to managers. On the other hand, he stayed too long in the job. When they merged PTD, Ramp, and D1C, the ones left in charge were the PTD engineers who were used to working in deadlines of months, not days and who always had extra resources to bring in the design process. When in 10nm they thought they had a solid process because they could run wafers though a path of golden tools, they ramped up and yield turn to sh*t because the rest of the fleet hadn’t been vetted. They just assumed the other tools would match.
Ramp engineers and line managers, who had never missed a commit through 1272 (starting in 1262), watched each tool to make sure they were ready for prime time - but they were forced to go else where in 2016 unless they wanted to work under PTD senior managers who didn’t understand line management and the absurd focus on hot box numbers and velocity. The CFO later reported that the hot box craze cost us 7% production and $500m in losses. Unknown is the quality loss because we let prod lots sit over queue time in preference of some HB2 engineering experiment.
In the end, we were probably sunk a long time ago. We just can’t compete with TSMC because they have much larger resources, economies and learning of scale, and a much cheaper labor force. If we had gone fabless 10 years ago, it might have been different, but at the point, we (the fab folks) wouldn’t be working for Intel anymore anyway.
It was under Sohail that PTD flourished. Key industry inflections happened under him (hik mg, finfet..). It’s precisely the lack of discipline and focus in the current crop (largely D1C plants) that LTD is in a ness. 18A is in shambles and there’s absolutely no denying of it - if this were false, why aren’t external customers lining up? Enough said.
sohail presided over a 10 year 10nm tech cycle, but he can bring back intel tmg? the reason tmg is the basket case it is today is sohail. anyone that works in tmg now knows that. the hangover of id--ts he promoted to vp is still ki-ling tmg.
He getting some H1B punani
Intel is infested with a bunch of yes men all over. Intel actually needs people who can firmly say no with all their might and conviction. Sohail was a great leader when Intel process ruled the world. His style of crack-the-whip discipline is whats needed to regain leadership. TSMC does not coddle its employees. Intel got too soft with psychological safety, WLB and job security.
Accurate. LTD is still paralyzed by this mentality - have recently had Sr. PEs wax poetically how we need him back and subsequently channel their inner SA when flexing by saying No. The stench of his cologne lingering in elevators and conference rooms after he'd been around is a core Intel memory for me.