fariszr

I agree. They are definitely trying to make the jump between 4o and 5 seem bigger, though most users never actually used o models, so maybe from a layman's perspective it was right to skip o1/o3/o4-mini?

GPT-5 is legitimately a big jump whe it comes to actually do things you ask it and nothing else. It predictable and matches Claude in tool calls while being cheaper.

GPT 4o ≠ GPT-4

The jump from gpt-1 to gpt-2 is massive, and it's only a one year difference! Then comes Davinci which is just insane, it's still good in these examples!

GPT-4 yaps way too much though, I don't remember it being like that.

It's interesting that they skipped 4o, it seems openai wants to position 4o as just gpt-4+ to make gpt-5 look better, even though in reality 4o was and still is a big deal, Voice mode is unbeatable!

The problem for me with dokploy is how do you manage in-config secrets for deployed apps. There are many apps which have secrets in their configs that can't be imported from env variables.

The only solution is to have these files locally on the server and then use an external bind mount.

I solved this in my docker-compose-gitops-action by just inserting secrets before copying the files to the server.

https://fariszr.com/docker-compose-gitops-github/

you can't do that with dokploy even with this dokploy action so it's triggers a pull from the source repo only.

Also the preview feature on dokploy is almost useless because there is no variable to get the preview URL dynamically.

What makes your service especially privacy friendly?

I think if you are striving for full privacy, you should implement the secure enclave idea presented by ollama, it makes the entire pipeline fully encrypted, I'm waiting for an actual provider to finally implement this.

https://ollama.com/blog/secureminions

That would only if they used the website not the phone app, which isn't really the case.

Reverse engineering.

Yeah but that means they have to be aware of the need to activate tailscale on their phones manually everytime they want to use your apps.

On PC I agree, you can just leave it running, on mobile though it chews through the battery like it's nothing.

This would work if they didn't use the apps, which in the case of nextcloud they do.

Great blog post, but unfortunately from my experience with my kinda tech-friendly family, i can tell you that not exposing service publicly is an absolute UX killer.

Nobody uses the local nextcloud because they just don't think they can rely on it, it doesn't always work from their perspective, and is too finicky to use, because it needs an external app (Tailscale).

This can be only fixed when the app itself can trigger a vpn connection, and I don't think this is going to happen any time soon.

This looks very impressive, but the Readme has way to many details, I think i got the idea but I'm not sure, and that's a problem.

APM looks interesting, would love to try it out.

AFAIK they used Gcloud to run their models.

This makes sense in the grand scheme of things. Anthropic used to be in the Google camp, but DeepMind seems to have picked up speed lately, with new “Experimental” Gemini Models beating everyone, while AWS doesn't have anything on the cutting edge of AI.

Hopefully this helps Anthropic to fix their abysmal rate limits.

I built something similar, a github action for doing gitops with docker compose (or swarm): https://github.com/FarisZR/docker-compose-gitops-action

For other stuff I just use another version of the action to deploy files using Tailscale SSH: https://github.com/FarisZR/tailscale-ssh-deploy

A lot of the efficiency improvments depend on the new thread scheduler in Windows 11.

I doubt these chips are going to be great or at least, as efficient in Linux as they are in Windows 11. AMD going for Zen5 and Zen5 is exactly to avoid this issue and the need for complex scheduling algorithms.

> Would be awesome if they provided a pre-built container for each module,

I created a project that does this, it automatically builds images on new releases of caddy or any of the plug-ins. It builds images for almost all caddy-dns plug-ins, thanks to unlimited github actions minutes :)

https://github.com/FarisZR/caddy-dns-OCI

This is probably cloud related, the photos were taken from an iPhone, yet resurfaced on an iPad, according to the user.

> The latest report suggests that the erased and sold iPad is somehow restoring old photos from an Apple ID that is no longer signed in to it. The Reddit user says the photos that are reappearing are from 2017, which is in line with similar reports. The images were initially taken on an iPhone, and so had been synced to the iPad via iCloud Photo Library before the iPad was wiped and sold.

This is probably iCloud related, I guess that's what you get for using unencrypted storage.

Ente is finally a realistic encrypted alternative for Google photos / iCloud if you are looking for one, it's also open source, but it's yet another subscription and it's not cheap at all.

> The new law sets a new standard: People should be able to use end-to-end encryption "wherever it is technically possible". In the text it is explained why this clear requirement to cloud providers is necessary: "Although end-to-end encryption is now the industry standard, individual messenger services do not use end-to-end encryption or only use it for certain functions, without this being justified by technical restrictions."

It's only mandatory where it's technically possible, Gmail, Instagram, Telegram and others are the ones likely to be affected.

Read some details in the issue and you will get an idea why they haven't migrated yet The size of the issues in their repo is huge! Especially when including media like photos and videos.

Also funding hosting for such a thing is challenging, AFAIK they found a Chinese sponsor, but some of the community weren't so pleased with them.

According to the article, AMD seems to have pulled the plug on this as they think it will hinder ROCMv6 adoption, which still btw only supports two consumer cards out of their entire line up[1]

1. https://www.phoronix.com/news/AMD-ROCm-6.0-Released

It really shows how neglected their software stack is, or at least how neglected this implementation is.

> after the CUDA back-end was around for years and after dropping OpenCL, Blender did add a Radeon HIP back-end... But the real kicker here is that using ZLUDA + CUDA back-end was slightly faster than the native Radeon HIP backend.

This is absolutely crazy.

> I will not know if the person I'm talking to is using the official Signal app or Tiktok, will I?

Well, if meta is going to sign an agreement with the app, it's probably going to show which app a user is using, especially to avoid username/phone number collisions between platforms.

> I think it was only a change in the ToS that was reflecting what they had already been doing in practice for years, right? Most people I know did not really care, they just got exposed to alternatives (Signal, Telegram, Threema), and then realized that WhatsApp was just fine for them.

Nope, it was much bigger than that, it also heavily widened the scope of the data they stored without encryption.

> I am not sure it will, for the same reason almost everybody uses Google Chrome (not a Chromium variant) or Safari even when there is no lock-in: people just use what they are exposed to.

Chromium isn't relevant here, Email is. Even though Gmail and Outlook control like 90% of the market, it didn't stop competitors like Proton or Hey from existing, it allows others to innovate without hitting the same issue, lack of users on their platforms.

> An open API enables third-party apps, and that's cool. But forcing WhatsApp to support third-party protocols in their own app is just weird to me

I don't think they have to support anything, it's just an open API, the other app has to be compatible with WhatsApp to be able to send anything. Also, why not? They have the capability to do so.

I find matrix's article on this very useful https://matrix.org/blog/2022/02/03/digital-markets-act-and-i...

> how do I know that I can trust the protocol?

It's going to show you which app and probably which protocol is going to be used, then you can decide on your own how to proceed.

> really fear that in practice, it will just weaken end-to-end encryption everywhere. And for what? I have absolutely no problem using multiple apps to talk with friends right now, I don't need one "super-app".

You always have the option to just use the same app? That won't go away, and it will still have advantages

> But maybe that is actually a political incentive: if you can't make E2EE illegal (in order to spy on your citizen), make it weak.

The idea is to kill the network effect in chat apps, it way too strong now, to the point that even the WhatsApp policy change controversy didn't kill it, everybody complained, opened a signal account and then just continued using WhatsApp

And of course they did the master class of withdrawing the change then rolling it out when the news cycle fizzled out, so that 'protest' had literally no effect on WhatsApp and it's dominant position.

I never use Twitter to keep up with anybody who I know in real life, Its always where I go to get the latest news, the latest updates about anything I am interested in and musk doesn't seem to really care about that.

Mastodon is way too niche to have the variety and the relevancy of X and Facebook is just way too cowardly to build a platform that's actually relevant to real life and not just wasting time while selling as many ads as possible.

I think the best way to be up-to-date with people is ironically to use messenger app stories, they are almost always encrypted and will show up to all of your contacts without any algorithms in between.

Honestly mastodon was never a replacement for Twitter and it's not even close to the variety Twitter has.On Twitter you can delve into pretty much anything and get the real time events about it whether its, sports, politics, economics or even tech and it's own niche cultures

Nothing come even close to that, mastodon is just a couple of small niches within tech and some news organizations, you won't find anything near to how Twitter works and you won't find anything even close to being relevant as it is.

I personally think most of the complaints that hacker News writes about Twitter, stem from who you follow.It really depends on who you follow, if you find the right people it's going to be your source for what's going on in anything you're interested in.

And unfortunately Elon doesn't seem to be interested in keeping that, he is just interested in milking every last dollar out of it and the result is the platform fighting against its own users.

Unfortunately he also proved something very dystopian while he's doing this:The network effect is so strong that even if you fight against your users، nothing is going to change.

This article explains it fully: https://stratechery.com/2023/threads-and-the-social-communic...

My comment on it is also relevant here: https://news.ycombinator.com/item?id=36684443

> the court ruled that the assignment of automatically calculated credit scores is not in line with the GDPR.

IMO The biggest point in the article, that's why I modified the title a bit.