I enjoy speculating on what could come from revisiting the first principles with our new tech stack.
The Fediverse makes it much easier to broadcast updates in a truly decentralized fashion. Maybe this can be the discovery layer instead of requiring a centralized social network in GitHub?
Git itself has many of the primitives needed, but issue tracking and CI/CD seem the main pieces still lacking. Seems likely that issues can be solved satisfactorily both natively in git, or in the fediverse. I bet that agents can pass branches/patches amongst themselves just fine, maybe native git (pulling from each others’ remotes on a crontab, or sharing patches by listserv) is actually completely viable now.
I also wonder if agents will cause a shift from forking and contributing back upstream, to more like horizontal gene transfer where people or tight teams broadcast their repos and other projects borrow/steal whatever may be useful. You see this with the Claws where maintainers are simply swamped with contributions and no way to verify whether they are correct/positive. (Things like “first time install on a Mac mini breaks at step 7, due to some iMessage issue” are fiendishly difficult to automate”)
It does seem like the USG should fund a few $10B to build out the alternative stack, given the strategic sensitivity. Best time to have started would have been 20y ago, second best is today.
The US Government made ASML dominant when it allowed it to acquire (US Company) Cymer, Inc.- the company that was best in the world at the time at EUV. Merging Cymer's EUV work with ASML's meticulous perfection and delivery of the entire rest of the system is what made them the only vendor that matters for semiconductor manufacturers.
This acquisition is also what gives the US Government the ability to veto customers of ASML even today- this is why Chinese semiconductor manufacturing is so far behind, because the USG controls who can access ASML's EUV work.
Run Facebook on a single Proxmox box and demand would still outstrip the supply.
What yet needs to be seen is if that demand sustains in the long run at that price point or flattens out proving to be super elastic given that there are many other providers that are catching up pretty fast.
This one seems clear cut as a HIPAA violation. Glad to hear that interpretation was upheld.
However, regardless, we really need to just kill the data broker business model.
Speaking as someone who implemented GDPR for my startup when the law first came into effect, there were certainly rough edges.
But the core premise that you simply cannot sell user data to sub-processors without consent is a powerful one that I believe would fix a lot of broken things in the US system.
(Not least because the USG buys private data that would be unconstitutional for it to directly collect, but also things like the incentives for your cell phone provider to sell your location data to advertisers.)
Seriously, we have a country where a large fraction of our ad spend is for services that promise to remove your private data from data brokers. We could literally just pass laws so companies could not do this.
TL;DW: HIPAA was actually created to allow insurance companies to share patient data without having to get patient consent. Before HIPAA, data was more fractured and less commonly shared. The only privacy protections it offers is, e.g., your doctor not giving your data to your boss. But about 1.5 million private entities can legally access your data (everything from health startups to insurance companies to hospitals)
> But about 1.5 million private entities can legally access your data
Somewhat. They are allowed to access it "for treatment purposes", not just to nose around out of curiosity.
I found myself explaining this to a number of my patients (I used to be a paramedic) who were irate about disclosures they'd made to their therapist, doctor, etc., that they had said they didn't want revealed to other providers (but were actually germane to their care).
"Does the HIPAA Privacy Rule permit doctors, nurses, and other health care providers to share patient health information for treatment purposes without the patient’s authorization? Answer: Yes. The Privacy Rule allows those doctors, nurses, hospitals, laboratory technicians, and other health care providers that are covered entities to use or disclose protected health information, such as X-rays, laboratory and pathology reports, diagnoses, and other medical information for treatment purposes without the patient’s authorization."
One problem is all the data breaches it encourages. Data breaches are already bad enough with the providers I actually use without 1000s of random companies having access.
Reminds me of this Seinfeld episode when Elaine was marked as "difficult" in her chart, and then she couldn't get a single doctor to see her. She wasn't allowed to see her chart or edit it after that. As soon as she got to a new clinic, they would receive a phone call from another doctor warning them not to treat her.
Yes, and - there is also something about the visceral feeling you get when your turn comes up in standup and you didn’t update any tasks and you don’t know the status of the thing you promised for this week.
If the PM does the task list and then chases the engineers 1:1, it’s a different chimpanzee brain mechanism at play. Very easy to forget/ignore you are letting down a whole team in this mode.
(And the flip side is true too, shared victory is more motivating.)
This is exactly why I do it (as a PM). The average engineer gets way more of the important work done when they have to say what they did and when, if they're blocked, they know I'll actually help unblock them.
And when the standup itself is five minutes, people are still refreshed enough to talk about a book or tv show or show off the progress they're making building a deck or let their kid say hi.
> there is also something about the visceral feeling you get when your turn comes up in standup and you didn’t update any tasks and you don’t know the status of the thing you promised for this week.
Never really experienced this. But daily are boring when it goes past the act of sharing updates and into musings by the PM, design discussions with a few of the team while the rest idle…
The West-German debate in the 70s came from the realization that the sheer size of the Holocaust/Shoah was in no small degree due to bureaucratic record keeping. Storing someone's ethnicity is potentially dangerous for that person.
TLDR, most of what is getting solved so far is “easy” problems that were not seriously looked at by experts, and where there isn’t a new insight, just trying all the existing techniques from the toolbox. Essentially the low hanging fruit for automation. Raw count solved is a problematic eval due to its difficulty lumpiness.
Seems this problem might be different, having some new insight as part of the solution.
reply