I built a bash script to manage deployments of multiple applications to a single Dokku server. I made this after I was paying like 30$+ a month on DigitalOcean, and wanted to move fast to a different VPS
What it does:
- Import existing servers - Run ./deploy.sh --import ./apps --ssh your-server to clone all your apps, export their config (domains, ports, storage, postgres, letsencrypt), and env vars
- Migrate between servers - Import from old server, change SSH config, deploy to new server
- Multi-app orchestration - One config.json manages multiple independent repos
- Hierarchical configuration - Parent settings cascade to deployments, children can override
- Smart deploys - Skips unchanged apps by comparing git commits
- Tag-based filtering - --tag staging or --tag api to deploy subsets
- Secrets via .env files - Hierarchical: _api (shared) → api.example.com (specific)
Pure bash, requires only jq. No daemon, no external services.
1) Surrogate endpoint only — HPV PCR positivity is not a clinical outcome; no CIN2/3, no cancer, no mortality measured
2) Correlation ≠ causation — HPV-cancer link is epidemiological association; Koch's postulates not fulfilled in traditional sense; detecting DNA doesn't prove pathogenic activity
3) PCR detection ≠ disease — Transient HPV infections are common and clear spontaneously; most HPV-positive women never develop lesions or cancer
4) Type replacement signal ignored — 66% higher incidence of non-vaccine HR types in vaccinated group is dismissed rather than investigated as potential clinical concern
5) No long-term clinical follow-up — Cervical cancer takes 15-30 years to develop; this 7-year study cannot assess actual cancer prevention
6) Confounding in vaxxed vs unvaxxed comparison — Unvaccinated group is small (n=859), likely differs in health behaviors, screening adherence, socioeconomic factors
7) Circular reasoning — Vaccine "works" because it reduces detection of the types it targets; says nothing about whether those types were actually causing disease in this population
8) Assumes HPV16/18 reduction = cancer reduction — Untested assumption; clinical benefit must be demonstrated, not inferred from PCR
9) High baseline HR-HPV in vaccinated group unexplained — 32% prevalence of other HR types suggests substantial ongoing oncogenic exposure despite vaccination
10) Genome validity unestablished — HPV reference genomes are in-silico constructs assembled computationally; never validated by sequencing purified, isolated viral particles; PCR/sequencing performed on mixed clinical samples where true origin of amplified fragments is indeterminate
> In the phase 3 trial, 296 patients with advanced forms of non-small cell lung cancer were randomly assigned to receive either lorlatinib (149 patients) or crizotinib (147 patients, of whom 142 ultimately received treatment).
Well said. If there's no true freedom of speech, we've lost everything.
The last two years offered already a glimpse... Now look at China. It's mind blowing. Poor people.
As time goes on, China will appear less authoritarian, and more desirable to people with a certain (warped one I won't name) ideology, as it continues towards its logical conclusion.
China's one child policy now seems "reasonable"
Their mass surveillance is almost desired now, with our new "Ministry of Truth"
Jail time for speaking out against the government -- something I saw plenty of people vouching for when dissenting views about the pandemic were offered up.
Re-education camps for speaking out against teaching certain ideologies and concepts to young children are almost implicitly implied.
There's more, but it's clear this is where we are headed. The slippery slope isn't a fallacy, and 1984 was a prediction.
You don't really have to trust anyone's opinion on this... You literally just need a yardstick and a calendar. Or a pair of eyes. Snowpack gets worse and worse every year.
> Or a pair of eyes. Snowpack gets worse and worse every year.
Yup. I spend a good deal of time training for and rescuing people in the Sierras. Snow pack is very obviously lower than a decade back. Sure, sure, we'll get the occasional outlier like that massive snow pack a few years back, but ever other year, it gets lower and lower.
I have no opinion about flying on this particular plane.
It's not so much MCAS, electric trim, this, or that.
The bigger issue is it's too complicated, has too many features, and the FAA can't adequately oversee testing and certification all the way down the engineering stack. Furthermore, because of pressures to save a buck, Boeing is willing to cut corners and sacrifice safety by slapping a plane together without properly engineering or testing it. Because of this behavior, it's difficult to know how many other problems are lurking around like in the 787.
This is every commercial plane, though. We've traded more frequent crashes due to human error for fewer, more spectacular crashes due to how complex planes have become.
False equivalency and hasty generalization that equivocates every plane as having the same risks when they clearly don't, and ignoring testing and training process improvements. Also, an oversimplification that somehow fundamentals of aviation are shifted as a finite resource over to automation when that's clearly not the case.
737 classics are simpler and reliable.
The NG's randomly have hidden structural weaknesses exposed during runway overruns and hard landings when the fuselage breaks up because of Ducommun and Boeing negligence.
The MAX is a steaming pile that may be a white elephant around Boeing's neck.
The 787 is notoriously-bad.
The 777 is pretty good.
Airbus has had much more automation for years. Are they dropping out of the sky? Maybe they manage their complexity better than Boeing, who hires underpaid engineers to work on things that they're really not qualified to do.
I have taken a similar stance myself (not that I'm flying anywhere these days) but, out of curiosity: do you also refuse to be a passenger in automobiles? The 737MAX is quite a bit safer than those.
The 737MAX is safer than a car per passenger mile, but saying that it's safer than "driving" is false unless you assume that both trips are the same distance. This isn't usually how it works. A Californian who refuses to vacation in New York to avoid the 737MAX usually does not drive to New York instead -- they drive to Point Reyes or something nearby, or just don't go at all.
This is true to some extent, although the difference in safety is so large that in the example you are still more likely to die from a car accident on your way to Point Reyes than your flight to New York.
Deaths from car accidents per 100 million miles driven:
1.33
Deaths from plane accidents per 100 million miles by plane: 0.0077
(Note above statistics are from the US)
In your example, California to New York would be approx 3k miles while point Reyes is approx 150-300 miles depending on where you start. A 10-20x longer journey via plane is still an order of magnitude safer.
Per mile, the car is about 200x more dangerous - in fact the usual statistic is that getting on a plane is safer than driving to the airport, not doing the whole distance in the car.
All of these statistics are too generalised to be useful.
People who fly in planes are generally not flying from every conceivable location on the globe to every other one in an even distribution and randomly using every type of plane available on every carrier. Neither are those driving cars driving from every possible destination to every other in every kind of vehicle etc.
Compare someone exclusively flying back and forth between one wealthy first world country to another on a Dreamliner run vs someone driving an 80's Ford Pinto regularly across treacherous mountain passes in the Andes.
Or the converse (use your imagination).
An important and useful statistic is that one particular, new model of aircraft is tremendously crashy considering the amount of time it's been in service.
It makes sense to avoid travelling on that aircraft, until it can be proven to have a flight-miles/crash ratio more in line with other models. (Which may take a number of years, or decades, especially if people are avoiding it...)
What you say is true but coarse statistics doesn't reflect actual situation. You would have to compare 737 MAX deaths per mile to OP car deaths per mile. Maybe taking into account that hopefully OP doesn't speed, drink, text, watch harry potter... Don't know if deaths includes pedestrian or bikes killed by cars, could be irrelevant if you only compare your odds to be alive after journey.
I'd also expect a recent car with good safety to have a lower rate than average and maybe 737 MAX higher than average?
Apologies, I thought OP was just talking about flying in general rather than specifically the 737 max but I have re-read and it's clear. I think I just forgot as I was writing the response!
> Don't know if deaths includes pedestrian or bikes killed by cars, could be irrelevant if you only compare your odds to be alive after journey.
I agree this could be debated, but I don't think it's too useful to say "The journey was safer because rather than kill the person travelling it only killed a random pedestrian".
I find the +/- sentiments interesting, it wasn't any intention of mine to approve/disprove anything, I was just sitting at home thinking were the original charts meaningful.
I built a bash script to manage deployments of multiple applications to a single Dokku server. I made this after I was paying like 30$+ a month on DigitalOcean, and wanted to move fast to a different VPS
What it does:
Pure bash, requires only jq. No daemon, no external services.