4 ms·

I think Spectre and related vulnerabilities are directly comparable to the 737 MAX debacle, insofar as they reflect poor engineering decisions made directly aga

I think Spectre and related vulnerabilities are directly comparable to the 737 MAX debacle, insofar as they reflect poor engineering decisions made directly against the customer in favor of short-term profit. Of course Boeing’s decision was far more devastating and led to the loss of 346 souls, but the pathology that gave rise to both situations (executive hubris, failure to tackle accumulated tech debt) seems quite similar.

ghaff 6y ago
Eh. Some of the vulnerabilities in that class were Intel-specific. But there is a whole class of vulnerabilities that potentially can hit any design that has speculative execution--and people just aren't OK with shutting off speculative execution because the performance hit is so big.
- ericbarrett 6y ago
  Intel chips were significantly more impacted by Spectre mitigations than AMD & others[0], and needed significantly more workarounds. [0] https://www.tomshardware.com/news/intel-amd-mitigations-performance-impact,39381.html https://www.tomshardware.com/news/intel-amd-mitigations-perf...
totalZero 6y ago
Workarounds for Meltdown and Spectre can impose a nontrivial performance hit for certain workloads, but it's tough to compare a 20-month grounding of 737 Max with patched security vulnerabilities that the average consumer neither understands nor directly observes.
- ericbarrett 6y ago
  Why not? Obviously the outcome of Boeing’s decision directly led to many deaths, and I am explicitly not trying to compare the two on result. However, the corporate actions which created these two scenarios are quite similar and I think it’s useful to compare them, since we’ll definitely see it again and again in the modern business world.
xxpor 6y ago
The thing about the max is the deaths were directly observable and obvious. The problem with Spectre/Meltdown is they might have been exploited for years, leading to wars/economic harm that the general public will never be able to tie back to Intel. On the balance of probabilities I think the Max debacle ended up doing more harm, but IMO Intel (and AMD/ARM to a lesser extent) did get off easy because of the extremely technical nature of the issue. Edit: I also agree that the max issue has a much more direct Executive Directive -> Harm line to draw. I don't think the Intel CEO went down to engineering and said anything comparable to "create a new version of a plane that needs no new training while having completely new larger more efficient engines, even if that's physically impossible"

ghaff 6y ago

Eh. Some of the vulnerabilities in that class were Intel-specific. But there is a whole class of vulnerabilities that potentially can hit any design that has speculative execution--and people just aren't OK with shutting off speculative execution because the performance hit is so big.

ericbarrett 6y ago
Intel chips were significantly more impacted by Spectre mitigations than AMD & others[0], and needed significantly more workarounds. [0] https://www.tomshardware.com/news/intel-amd-mitigations-performance-impact,39381.html https://www.tomshardware.com/news/intel-amd-mitigations-perf...

totalZero 6y ago

Workarounds for Meltdown and Spectre can impose a nontrivial performance hit for certain workloads, but it's tough to compare a 20-month grounding of 737 Max with patched security vulnerabilities that the average consumer neither understands nor directly observes.

ericbarrett 6y ago
Why not? Obviously the outcome of Boeing’s decision directly led to many deaths, and I am explicitly not trying to compare the two on result. However, the corporate actions which created these two scenarios are quite similar and I think it’s useful to compare them, since we’ll definitely see it again and again in the modern business world.

xxpor 6y ago

The thing about the max is the deaths were directly observable and obvious. The problem with Spectre/Meltdown is they might have been exploited for years, leading to wars/economic harm that the general public will never be able to tie back to Intel. On the balance of probabilities I think the Max debacle ended up doing more harm, but IMO Intel (and AMD/ARM to a lesser extent) did get off easy because of the extremely technical nature of the issue. Edit: I also agree that the max issue has a much more direct Executive Directive -> Harm line to draw. I don't think the Intel CEO went down to engineering and said anything comparable to "create a new version of a plane that needs no new training while having completely new larger more efficient engines, even if that's physically impossible"