I don’t consider myself progressive. But patents are a critical part of innovation.
Glad your industry pays you. In mine, pharma, the innovating scientists usually get $1
That brings back the memories. We used to get £1.
j
The Newton was a John Sculley project. Jobs disliked it and killed it upon his return.
Really? I mean, a salary would be nice.
Fair enough, but he clearly saw the potential. They went big on similar devices later, when technology caught up. Though without the stylus, which Jobs hated literally to his dying day.
It’s interesting, SpaceX has generated a lot of patents that don’t list Musk (note that this is an assignee list, but some random checking does find SpaceX employees).
Musk himself as inventor is on a narrower list.
Nope. Just $1. No patent, no lunch.
Joking aside, of course there is a salary. But every company I’ve worked for (and it’s several at this point) makes you sign something when you join that says your inventions are theirs, and you will actively help them to get the patents. I was told (and assume is true) that the $1 satisfied some requirement to consumate the deal.
And I’m not a scientist. I’ve been part of the bureaucracy my whole career, but I still have to sign.
Wow. My company paid $3000 if your patent application reached a certain point, which ours did, another $5000 if it was granted, which ours wasn’t. Which was a pretty strong incentive to try to put something interesting together that our lawyers thought stood a reasonable chance of being granted.
It’s to satisfy the “consideration” portion of the contract. (Enforceable contracts require an exchange of promises. You promise to give them your patents, they promise to pay you a dollar) It could be handled as part of the salary portion of employment, but it’s an easy and safe way for a timid lawyer to draft the agreement.
Actually, that makes a lot of sense. Thanks
Looks like there’s a recurrence of January’s screwed-up attempt to tighten the Twitter API. Currently getting this hitting what’s supposed to be the web endpoint:
What’s notable about this is that it’s coming off the web port, and it’s lasted long enough for me to notice. I’ve seen lots of glitches and errors since 11-2022, some never being fixed, but this is the first time I’ve seen an actual outage lasting this long.
I expect this will get fixed in an hour or so, but it’s notable that we’ve reached the point where this kind of rookie error can reach production, and not get fixed in ~30 mins or so. This is a bigger issue than somebody simply neglected to pay a bill; it points to a significant deficit in QC and SRE capabilities (loss of staff, institutional knowledge, diligence, take your pick).
Ultimate agility.
ok, fixed now after a little over an hour-long outage.
This doesn’t mean Twitter is imminently crashing, but it does mean that it’s reached the point where there are inadequate safeguards protecting against this kind of issue, and it takes them over an hour to figure out how to revert the change.
This is a pretty remarkable milestone for a company of this stature. When they hit a real problem (infrastructure failure or critical database corruption), it’s not going to be pretty.
My mention of Steve Jobs and the iPhone was intended – before the discussion devolved into a hijack about patents – to point how how intimately involved he was with the development. He obviously didn’t personally design all its electronics, but he envisioned a revolutionary novel product, knew exactly what he wanted it to be, and relentlessly worked with the iPhone team to make every detail of his vision become reality, down to the last detail about which Jobs was known to be obsessively meticulous.
Pretty sure I must have got this impression from the Walter Isaacson bio of Jobs, which is a great read, and shows this to be Jobs’ typical work method.
OTOH, I got no such impression from the Ashlee Vance bio of Musk, although it’s a different author with a different style. But the impression I’ve built over the years from multiple sources is that of an overbearing micromanager who hires good people, meddles in their work, and fires them if they don’t agree with him. I’m sure it’s no coincidence that (except for the “hiring” part) this is exactly what we’re seeing at Twitter. And I’ve told the story before of Elmo’s personal assistant of many years approaching him with the observation that as the size and complexity of his enterprises grew, she was taking on more and more responsibilities, and that this should be reflected in a higher salary. Thinking this over in a way that only Elmo could, he fired her.
Incidentally, the respected biographer Walter Isaacson is nearing completion of his own book on Musk. It will be interesting to see the perspective that he’ll offer. Reports are, however, that Elmo’s wild antics are making it hard for him to finish the book. Every time Isaacson feels he’s bringing it to a satisfactory close, Elmo does something newly outrageous again.
Interestingly, it sounds as if Twitter under Musk is heading the opposite direction from Agile practices.
Whatever you think of Agile, it does have the benefit of shipping smaller, more frequent changes in order to reduce the blast radius of errors (i.e. move fast and break things).
From what I’m able to piece together from inside sources, Musk won’t tolerate the “break things” part of that equation, so instead they’re implementing continuous code freezes except for changes he’s personally authorized. This means batching up more changes to be delivered at once, meaning a broader scope of change for deployments, meaning a higher likelihood of missed test cases and higher effort to reason through what actually caused the breakage.
In other words they’re headed away from Agile and back into waterfall methodology. This is exactly what you’d expect from someone who last coded in late 1990s, or whose risk calculus is calibrated with a sensitivity that’s more appropriate to enterprises where the cost of failure is much more painful – say, a vehicle recall or a rocket blowing up on the launch pad, or a narcissist CEO getting roasted on social media. In a bid to avoid these outcomes at all cost, you get into a spiral where risk avoidance leads to more failure, which leads to more risk avoidance, and so on.
If this goes on, I think it’s reasonable to expect a slower pace of feature releases, and deployments scheduled on weekends rather than continuously, occasionally resulting in outages that run the course of the weekend and start spilling over to the weekday. This would bring Twitter to the state of the art circa 2004.
I’d expect this kind of error would be noticed during the rollout (even if no testing was done in staging). Rollouts are supposed to be done a few replicas at a time specifically so that if the updated replicas start serving unexpected errors the rollout will be aborted automatically without very many people noticing. For this level of error, the start->abort time should be less than a minute.
I assume Twitter of old was doing proper canarying. Maybe they took it down to push a new feature (maybe the API permissions in general).
That’s what I would expect if this were a software issue, but I don’t think that’s what this was. It seemed like the web site was redirecting to the API host which may have been a DNS/load-balancing issue.
That’s definitely the sort of thing that can take an hour to fix if you’ve fired everybody except for “people who can write code” (which reportedly has been Musk’s approach). SREs don’t (generally) code, but those are quite literally the last people you want to fire because they understand the infrastructure best.
I used to have a T-shit with the same quote that the World’s Most Interesting Man is saying.