I mean, I agree, but whenever it comes up there's always someone who refuses to call a spade a spade... I'm curious how the pro-prediction market set would argue this is ok (other than to shrug and say "it's just one more thing to predict, why would we try to decide what can and can't be predicted?").
Don't rely on Google Takeouts for byte-perfect backups of your Google Photos library.
Depending on the device that uploaded the media and your country, they may have stripped GPS metadata from your media. I've also had captured-at times get changed randomly (!!) from your original media.
If possible, pull your media directly off your smartphone to make sure there are no shenanigans going on.
(These were painful discoveries made while trying to help support my PhotoStructure users)
Just to clarify, I've open sourced a ton of PhotoStructure (exiftool-vendored, batch-cluster, fs-metadata, @photostructure/sqlite, sqlite-vec, and 10 more repos), but PhotoStructure's frontend and backend integrations are not open source. If PhotoStructure, Inc. ever closes, though, I will open source the repos powering the consumer product.
I've written open source for several decades, and certainly enjoyed the spoils of open source, but I've also seen a ton of bad behavior. Keeping it closed source certainly is not a guarantee that I avoid intellectual property theft or license violations, but at least I'm not just one agent prompt away from someone rebranding and reselling what I've worked full-time on for more than a decade.
Open source let your project be rebranded and resold. That happened to me a decade ago. A company named ClickUp rebranded https://github.com/photo/frontend. It’s a risk with Apache 2.0.
Turned out not to materially impact our efforts. But it felt like a gut punch at the time.
AI enables this for closed source. Reputation and community are durable differentiators nowadays. I hope your 10 years of investment provides that.
This is what I did, but I found that it didn't preserve metadata when I tried to reupload my photos to Immich... everything showed up as though the picture was taken the day I uploaded it.
Last I checked (6 month ago or so) rclone was only able to download Google photos that had been uploaded via API/rclone. Other google photos weren't accessible.
But by using Claude Code, you're essentially handing over your data from Google's bot to another data-hungry bot run by an equally untrustworthy entity.
You don't need an anthropomorphized computer program running on trillion dollar companies' servers just to run `rclone copy`.
You're not syncing photos to anthropic services, you're not even using Claude code to run rclone for you, and specifically you're not ingesting every synchronised photo or video into the context.
You're not authorising CC to access your photos.
Claude Code is used to prepare the rclone run.
What do you mean that by using cc you're handing over your data (presumably synced photos) you're handing over your data"?
The comment I was responding to only said "just use claude code." So every single word you wrote in your reply is a hallucination.
If you genuinely believe that "just use claude code" means any of the things you wrote, you are definitely not careful enough to take the precautions necessary to contain commercial spyware. Claude Code would read your credentials and probe your files to "see if everything is set up correctly" and you won't think twice about it.
> OK, explain, what to the "just use Claude code" means in your opinion.
I see you didn't get the message. It's not about opinions. Don't add a bunch of extra meanings to suit your argument.
The original comment was:
> just use claude code to use rclone on a server and let it run automatically, its great and easy
First, it says "use claude code to use rclone." It's very simple English, yet it directly contradicts your claim that "you're not even using Claude code to run rclone for you."
> Because that's the only way I'd understand someone to go on such rant.
Yeah, no wonder given the above.
Going back to the original comment, it says it's easy. Easier than not using Claude Code, which is typing `rclone copy <source> .`. Every justification that you pulled out of thin air is more complicated than that.
> Is this what you'd do?
Again, you're not reading. Read my original comment. And the reply to your comment. I wouldn't use Claude Code in this situation. I'd just type `rclone copy <source> .`.
Right, I see, so when there's an opportunity for a rant you get one specific, narrow meaning, discarding anything else, even if more relevant and sensible, and then you go off.
Maybe try Claude Code one day, then you'll recognise what that phrase was actually about.
All the best! You must need it with such attitude to life.
You literally made a bunch of things up and directly contradicted the statement you so vehemently defended. That's not "sensible," and pointing that out is very different from word lawyering. No amount of "trying Claude Code" will compensate for that.
Just calm down. You seem to be personally offended at the mere suggestion that Claude Code isn't a good fit for everything, or that it's not even more trustworthy than Google. That's just a fact, not a "rant." If you feel the impulse to throw a tantrum because of that, I genuinely urge you to revisit your relationship with your cloud subscription because it's definitely not healthy for you, or the people around you.
That's on my someday-list. Immich through its rootless-docker configuration impressed me, but I've still gotta get around to figuring how I can provide access to family members. (I mean, I could buy some suitable hardware that isn't my desktop, but prices these days, amirite?)
You don't really need a new homelab, a used mini pc will do just fine. For me Immich idles at 0.1% on consumer CPU consuming less than 3GB of RAM:
- server: 1.40GB
- postgresql: 600MB
- machine learning: 125MB
- valkey: 25MB
This is for around 200k photos and videos. Mounting the thumbnails folder on an SSD should also help with load speed
Immich is not heavy. I bought an old mini pc from a second hand government office supplies website and it runs all my self hosted apps perfect with tons of room to spare.
I think the main thing is storage speed matters which is why a mini pc with an ssd runs circles around a raspberry pi running off a microsd card.
I haven't found setting up Immich to be super easy. But I have a special setup at home (Windows Server with minikube). I've been wanting to try it for some time though.
Maybe I should rope some NAS for just the Immich setup.
Windows Server with minikube seems like an unusual combination. I am guessing you would have a better experience self-hosting stuff with Linux (Debian) & Docker. Running Immich on that is fairly painless. Once you're familiar with that, then decide if something like k8s is right for you. Unless you're already happy with your minikube setup for other apps, and there was something specific you ran into with Immich?
I wrote the open source library that Immich (and PhotoStructure and several thousand other projects) use for metadata reads and writes.
PhotoStructure runs everywhere (laptops, desktops, macOS, Windows and Linux), libraries are portable across machines and OSes, there's a generous, usable free tier, and setup takes about three minutes. If you have any questions hop into our discord, we're a friendly bunch.
Not my experience at all. As part of my job, I’ve used takeout to export the data from hundreds of accounts, with total sizes ranging from <1GB to >500GB, without any issues. I genuinely can’t remember a time that it has failed.
I have nearly 600GB of photos in it and regularly do Takeouts every year or so. It takes a while and the format isn't that great (it's a ton of separate zip files), but it does eventually work.
My 2nd generation iPod stopped working last week. It had a spin wheel but its predecessor's actually spun.
The monochrome screen shows a folder icon with a warning sign. I haven't looked into it but I fear that something happened to the internal hard drive. It's one with spinning disks. It was truly a work of art, that device.
The HDD connector usually unseats slightly and it stops working. Pop it open with a credit card and reseat it. A more dangerous way that sometimes gets it working again is to wait for it not to be spinning and give the edge a good smack against palm of your hand. This usually works to reset it the tiny bit it needs…
The mechanical drive has probably failed. If you put the iPod in a freezer bag and get it very cold, it may spin up long enough for you to make a backup. But make sure it is in a freezer bag or something otherwise airtight or the condensation will brick it worse. The cold shrinks the metal enough to mitigate the initial friction of the drive spinning up.
From there, you can mod it to work from a micro SD card.
I’m sorry to hear that your iPod might be bricked. I haven’t worked on an iPod gen2 personally, but I repaired an iPod mini. The interface for the small hard drive in that is the same as compact flash, so it’s very easy to replace the storage with something more physically durable. Or at least it is easy to find the replacement part - I haven’t tried the whole process on your iPod.
Honestly I'm amazed yours lasted this long. I had a fourth-gen 40GB iPod back in The Day and it was easily the least-reliable Apple product I ever owned due to that internal 1.8" HDD. Mine was replaced five times under AppleCare.
When that eventually ran out and it died again (in 2009 IIRC), I swapped in a CompactFlash card using a cheap adapter from DealExtreme. Once you do that or an SD card mod (more popular these days), have iTunes restore the partitioning scheme and OS and it'll be good to go.
The distribution and customer research of building a durable and successful product and company is incredibly complex. LLMs can probably do each of the parts but it requires a lot of coordination.
Agent swarms, you say? Perhaps. But I don't think we're there quite yet. At least, we're not seeing it manifest.
I think it's difficult to tax people on money they've earned and "paid" taxes on. The billionaire tax in CA is a good example. It makes sense that the ultra wealthy would rather move than to pay that tax - for lots of reasons ranging from greed to wanting to decide how their wealth gets used.
It reminds me of student loan forgiveness. It's too hard to retroactively forgive them. It may sound good in theory but it's not fair - i.e. for people who worked 2 jobs while going through college.
The laws for wealth accumulation moving forward need to be reformed and I think we have to accept that we let people run away with massive wealth and it turned out to be not the best thing for society.
> it's difficult to tax people on money they've earned and "paid" taxes on
We already have property tax across the country. Sales tax is charged on post-tax spending. Hell, one could argue income is taxed multiple times given the SALT exemption.
Somehow multiple taxation is only an issue for inheritance and the rich.
> Imagine tax reform which collects an extra 1% of sales tax for items you have already purchased
Literally property tax.
> I just don't know how to get there
Constitutionally authorize a national property tax that can only be increased by both an act of the Congress and affirmation in referendum alongside the next Presidential election. (It’s high time we brought plebiscite into our federal system.)
Politically: use its proceeds to cut taxes for the middle class. Double the income tax brackets. Provide a per-capita citizen’s dividend.
I understand what you're saying. What I was saying is difficult is introducing a new tax that would tax already taxed dollars. It's the "new" part - but I get that most new tax laws will do this.
> What I was saying is difficult is introducing a new tax that would tax already taxed dollars
But that’s empirically untrue. Every sales tax is precisely this. Every property-tax increase, too. They’re unpopular, as are all tax increases. But I don’t think they’re more difficult than raising first-pass income taxes.
If the HQs move then along with them moves the high paying jobs and retail, housing, etc taxes also go. Rank and file employees at Meta can't dodge taxes as efficiently as their CEO.
i 100% doubt, one guy could disrupt the SV ecosystem - they might threaten to leave but they'll come back. The network and culture of SV which made these guys insanely rich is not something a few people moving away can disrupt. The rank and file might move to miami but the talented people will just find other jobs.
Elon moved the Tesla HQ to Austin which employs around 22k people. I'm not sure how much that moves the needle but it's not 0. California is doing just fine but if I were a public servant there then I'd think twice before making sweeping reforms.
my big issue isn't even the taxes, it's the fact that the money could easily disappear in bureaucracy even if it is collected. smh - the whole system needs a reboot.
People used to put effort into politics and government reform. But now we have the much easier/simplistic calls for a'magical reboot' on both sides of the isle.
The tooth fairy doesn't exist and paradise isn't a (very vague) societal reboot away.
My kids and their friends all dislike AI. I haven't put my finger on why, yet.
I think a part of it is a growing negative perception of technology. I think they're smarter than we might give them credit for and can feel when they're being exploited.
Maybe, despite not liking schoolwork they dislike more being a "dumb pipe" between AI and their teacher. We've stripped away purpose in the name of efficiency (and profit).
Some of the adults in charge....really shouldn't be. And I'm not referring to teachers or school administrators here.
My Frame TV developed a dim strip in the centre of the screen after four years of ownership.
Samsung would only proceed with organising a repair if I updated the firmware to the latest version. Something I deliberately held back on to avoid the introduction of ads.
Now I have new firmware. And new ads.
Oy for Samsung to tell me that spares are no longer available for my four year old TV, and would I like to buy a new one (with more ads).
Now everywhere I look I see Samsung displays being used in residential and commercial settings, and I'd estimate 30% of them have noticeably inconsistent backlights, with bright or dark spots.
I was using Claude to bounce ideas for parenting with a focus on biblical principals. Trying to really understand how to discipline my children without provoking them, for example.
I was pleased with the experience but Claude did not make it easy for me to dive deeper into scripture. I wanted a different LLM chatbot experience that was focused on less chatting and more scripture reading.
I TestFlighted it out and friends liked it so I ended up publishing it.
reply