Model X towing a semi uphill is fun to watch, but I think it only works because it’s on ice. Still, another vehicle would have gone nowhere https://t.co/ra5Wfe6wBa
Well said. I also think that the target in a step exponential is useless, so Q1 may well end up above 5,000 despite the lower target https://t.co/BQ30EUDOfN
Tesla has single-handedly created a revolution in transportation, and the substance of the revolution is the desire to find a much better solution to every problem and to attack obstacles head-on, something this type of article misses entirely https://t.co/LWQOHEPkCS
@SQLMaria@vldbb Thank you. There appears to be a coarse grained mutex protecting spm histories, so when adaptive is on with complex queries you inevitably create a huge bottleneck (cursor pin ...) I got around by fixing the baselines so plan histories don’t grow.
I think the real point of the Tesla Q4 Deliveries report is that @elonmusk knows how to efficiently scale the building process but not the quality control, and it makes sense, if a robot is moving faster than you can see and anything goes wrong ...
Musk thinks like (and is) a developer. He has his own goals and his view of the state of a project is determined by that. Like when a developer says “it’s basically done.” He’s happy now because now he believes they will actually reach 5,000/wk some day https://t.co/SJbRfEtCZp
@elonmusk I think it would be ok to miss an estimate that is six months away if you publicized step changes, then people would believe the S curve. And you could delay them by two weeks, just like stock quotes are delayed by 20 minutes.
@SQLMaria feature request: SELECT ... FROM ... ONCE, to read rows only once, would fail on attempts to re-read rows but not on snapshot too old, a lot of data migrations would benefit