> The Minab site, which was cataloged as an Islamic Revolutionary Guard Corps facility due to outdated data, was fed into Maven with other candidates and came out as a recommended day-one target. Target-list work that once took hours was condensed into minutes.
I feel like we're optimizing the wrong metric here.
Even ignoring the risk of civilian casualties (which to be clear should be weighed heavily) it’s just bad ROI. Like each missile probably costs more than the extra targeting work they should have done.
Yeah, totally, I’m being cautious because I don’t know how large the team is that would spend a few more hours, but under most assumptions you’re right.
Well you see you don’t understand. We’re America, the biggest and best military like no one’s ever seen.
We have the largest stocks of missiles ever. We can build them faster than we use them and cheaper than anyone else. So you’re a liar saying it would have saved money doing it any other way
I have found a lot of value in the approach the post describes. I store my markdown as GitHub issues, regardless of how I come up with a detailed feature description (usually a chat session).
Before implementation, context is always cleared to ensure that the model is working only from known context (skills, AGENT.md, etc).
A major benefit of this is that agents can look up prior prompt instructions as needed or when mentioned in the current prompt it's working on.
I know this post is about Noodle, but I want to talk about Immich.
I believe I've spent, perhaps, the most time of anyone on earth on digital photo management (hard to quantify, but since 2005 - 2 startups, one acquisition, and an ongoing open source project used by thousands).
I say that because I've refused to settle for most solutions in this space. Even when I adopted using Google Photos, it was as a read only viewer of my canonical photo library (I wasn't about to let Google take that responsibility). I have such high demands of whatever software I use for managing my photos and videos - they're really the only digital files I actually care about.
Immich is absolutely wonderful. It may have some shortcomings ... like partner sharing and sharing facial recognition between users. But man is it remarkable that an open source project can rival something from Google in terms of quality and experience.
And a soft fork like Noodle is precisely the way to handle it. I am perfectly happy with Immich so am not a user of Noodle. But it's open source flexing its strength.
I have great things to say as well. I left Google Photos about 1 year ago or two, used Takeout to export my ~terabyte of Google Photos, imported to Immich on my mac mini, along with more personal archives of photos. I store the images on an external SSD array (https://paulstamatiou.com/gear/owc-express-4m2-nvme-ssd-encl...) that I have backed up with Restic+Backrest to some cloud providers.
My mac mini hosts it behind tailscale to all my devices and it's been working great.
As someone looking to exfiltrate my photos from Google, I'm about to trial Immich, and I really appreciate you sharing your experience. Sounds like it's a good next step for me.
There's plenty of tutorials but you can request a Google Takeout of your photos and then point immich-go at it. I can confirm that this works really well.
Agree - it's annoying. rclone to the rescue. Requires choosing Google Drive as the destination and then downloading with rclone to your computer. It's the closest to sane I've found.
Claude made a PHP script for me that did API to Google Photos and pulls them all to my archive location.
I struggled with Immich, I didn't like the rclone option. I'm intending to go back to Immich but my first priority was to get the stuff out of Google. That's done, so with all my free time (ha!) I can get the family Immich going.
Not sure if it's still the same but years ago when you get your own photos from Google photos via it's APIs you could only get large thumbnails so a great quality reduction. So make sure yours are the original quality. Takeout was the only way out your photos in original quality.
I was planning on having Immich read from my archive. Then Immich on the phones sends right there (VPN?) and skips G all together. But maybe I'll end up with Amdroid->Google->PHP->Immich
Not sure if what I said made sense. I mean inject titles, descriptions, etc from the Google Takeout into the photos EXIF. Then when you import them into Immich, it'll extract those into its own database.
Since we're on Immich, please let me vent out my frustrations after having looked at it recently.
Why the love affair with forcing absolutely everything into strict temporal constraints? Is there a single view in Immich that isn't peppered with date headings?
And I genuinely do not understand what the sharing feature is even for. What are you sharing? It can't be albums because an album would be a series of photos you yourself arranged, independent of when the photos were taken, and this seems a completely foreign concept in Immich land.
My photos and videos are the most important files I have in my self-hosted storage system. Being end-to-end encrypted buys some peace of mind knowing that it's not end of world if my homelab is compromised. And I can backup the data to whatever public clouds without worrying too much about leaking data.
>But man is it remarkable that an open source project can rival something from Google in terms of quality and experience.
I'd argue Google is always way worse than the open source alternative in both quality and ux. Their products are terrible. From Gmail to Youtube, the product itself comes second (or Nth), while in open source the product is always first.
In my experience, most open source photo apps lacked basic features like facial recognition, mobile apps, etc. And where they do exist, the UX isn't the best. Are there ones I'm unaware of?
Kano predicted that users' perceptions of satisfaction with a feature will shift from delight to expectation over time. This is either because they have got used to it or because competitors have started to include it in their offerings. In the case of the touchscreen, for instance, the newness that came with the iPhone is no longer a novelty; the touchscreen has become the norm. It's no big deal anymore, but taking it away would be a big deal!
I’ve been a paying customer of Dropbox’s consumer plan for over a decade. Each time I visit their website I am greeted with a full page modal ad asking me to upgrade to a different plan I have no use for.
I’ve dismissed it every time. I don’t mind being shown that ad once. But it bothers me that they don’t care enough about my experience using their product to provide a “never show this again” option.
Or just use some metrics to know that if I dismiss it 50 times then maybe I’m not interested.
This is a bigger problem than it may appear. The business lifecycle is well understood, whereby a company innovates, expands, and then when it has a sufficiently large customer base, it monetises. I.e. it enshitifies its products - making them more expensive over time relative to the value. This process is death by a thousand cuts. A PM might see that having the modal you hate show every time you load the site increases conversion by 3%. That's $1M ARR you just made the company! Wow! You get a bonus and a promotion! The problem is you erode customer sentiment by some undefined unit. Many people think NPS measures this, but it doesn't. That measures how many people love and hate the product. It omits the vast majority of people in the middle. They just add it to the increasing list of annoyances until one day there are so many annoyances that they leave. At that time it's very difficult for the company to understand why the user churned. It wasn't any one thing. It was a concerted effort over years to squeeze every cent out of the user at minimum cost.
> At that time it's very difficult for the company to understand why the user churned. It wasn't any one thing.
It's very easy to know for people inside the company who actually care about the product and for their clients. It is really hard for directors and VP that anything they understand are numbers.
If I say "hey this modal/pop up is making customers angry" they are going to ask how much they care (is there even an unit for that?) and how much revenue is lost. The only way for them to handle a project is creating a goal for next Q and assigning people, when in most cases what you need is actually the opposite.
The problem is not that no one knows how to fix it, the problem is that managers organize around big problems and become useless at improving the small things or details. They become blind. It happened to me 100 times. They do not want things to be fixed unless there's a clear revenue increase, which is stupid.
I'm not sure why, exactly. But I don't pay any attention to news like this from Google. I don't know if there's some marketing which has me writing them off or if it's something else.
What I do know is that the Gemini integration into sheets is surprisingly incapable of performing basic tasks. This is where I expect Google to really shine. I expected Sheets + Gemini to be magical like Google Photos was. I hardly try anymore besides some basic math questions when I don't feel like inputting the formula myself.
The other thing I know is Google's propensity to sunset products. For many things, it's not a huge deal. And it may not be for this. But, why? When there are alternatives - both open and closed.
Not sure if this is connected, but I've found Gemini pretty hopeless within its own notebooks. I've been doing some job applications, and added cv and docs into a notebook. I then create a new chat to say "here is a job description, help me write a cover letter" or some such.
After about 3 messages in any given chat, a follow-up to "rewrite that with a more friendly tone" will result in a letter for a completely different job from another chat within the notebook.
On your Sheets + Gemini integration point, I've genuinely tried to give the Gemini integration into Google Docs & Google Sheets a chance. It is so incompetent that it is fully useless to me. I have not gotten a single correct solution each time I tried to use it, even something I consider table stakes. I often write my work reports in Vim in Markdown format, but they need to go to the corporate Google space. No matter how hard I tried, no matter how many prompts I have, it was completely unable to manage the command to "convert the Markdown format markers into native Google Docs markers". And I want to note, this was 2 pages of extremely simple Markdown with no "advanced" patterns, like tables or quotes, I think all I used was heading-marks, bolding, italicizing, and code blocks. This is something I would expect even GPT 3.5 to succeed in, and even more so Luna, but somehow it destroyed the formatting throughout half the document. This leads me to believe that they apply the absolute cheapest model they have there, or they have the model a harness which can barely be considered working.
I found it absurd when I found out that they suddenly made this Gemini integration an additional paid plan recently, there's absolutely no way I can consider that in good faith.
sorry if this isn't it. there is a hidden global setting that defaults to off that lets docs play nice with markdown. it's in file > settings i think. super annoying even if this is no help
I agree with the article - I don’t see how exposing children to AI in school ends well.
I don’t think we should expose developing minds to much tat we haven’t had a lot of time exposure understanding the impacts of coughsocial mediacough. The downsides just aren’t material.
Thankfully kids seem to already be jaded towards AI. Perhaps that buys them time to let their brains develope a bit more.
reply