miyuru

Google also, sometimes hides the search results it used for AI overview.

meaning you won't get them in the results page, but hidden inside the AI overview as links.

this and the increasing bad scammy ads, there is really good chance for other search engines to take over.

I sometimes watch those walking videos after visiting the city. Its whole different feeling than watching random video.

>Because the third-party runners sit outside GitHub’s network, they rely on a direct IP link to reach GitHub. The provider traced the hangs to intermittent degradation on that link.

This is interesting. If I read that correctly this means the CI host had a direct peering connection with Github or Azure network.

Why were they not able solve the issue?

the blog mentions that they use a local cache to "reduce the time" but the problem is not solved right?

Same here. It’s the first AI provider I actually gave money to, since they offered the model for free with a Mimo code for the first month or so, and it was great.

These days, there are more intelligent models like DS4.1, but Mimo is very obedient, so I plan things with another model and give the implementation to Mimo.

I had the same feeling yesterday, when the front page was filled with all apple stuff.

I have been lurking in lobste even though i dont have account, which currently has this post which is right up my ally.

https://lobste.rs/s/twkkac/fastly_speedtest_test

same thing on HN: https://news.ycombinator.com/item?id=49650630

take screenshots of the responses and do a chargeback.

I am working on a IPv6 only monitoring tool for the past 2 years.

https://v6monitor.com/domains/history/79/news-ycombinator-co...

Since I am working on other projects, its not fully what I want it to be ATM.

artificialanalysis for me.

https://artificialanalysis.ai/#intelligence-comparison-tabs

your dns server is down.

;; communications error to 2001:470:4b:27e:8::#53: timed out

same here.

Also in my country there is trend of remaking the old songs, with AI.

modernised sound with AI voice singing and AI video. They have ton of views in YT.

when trying to download models using browser, the DL domain for me us.aws.cdn.hf.co which points to bunch of singapore AWS IPv4 for me, despite it saying US.

Judging by the openrouter leaderboard ranking for today, it looks like Dv4F us more popular than mimov2.5.

https://openrouter.ai/rankings?view=day#leaderboard-table

These days cost per task is more important, and SOTA models have become expensive.

For me Cloudflare is worse, it takes more than 5 seconds, where as anubius take 1-2 secs.

funny with all the IP information they have, cloudflare cannot do a better job. (I am on IPv6)

and most of the time, its on marketing product pages like in framework main site, which can be cached.

I am from Sri Lanka, which is a large island.

We have a smaller number of ISPs due to the cost of submarine cables, and ISP prices were high due to profit-seeking. After Starlink came, the incumbent ISPs started to offer unlimited packages for the first time.

Also, Starlink is good as a backup connection for rural areas too.

> No, you can ask your provider to announce the IPs for you via a private ASN, but I have not seen a provider that is cheap enough to do BYOIP without an ASN. Getting your own ASN is the best option.

BYOIP in AWS is free. and VPC have lot of customisable and all are well documented.

https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-byoi...

Any plans to add IPv6 support? NAT GW are costly and docker hub has full support.

currently reg.mini.dev does not have AAAA records. Did not check the blob storage endpoints.

internet is used by billions of people, not just you.

didn't Scaleway also increase prices 1st of this month?

https://news.ycombinator.com/item?id=47944914

> i'm currently using their hillsboro instances.

existing VPS will have the old price.

It get increased only if it is rescaled and for new VPSs.

yes, I am using mimo code(free version) for the last 2 days. I gets the job done for me.

If I need to upgrade, the plan start at $6, so its a no brainer.

it looks like all AI is following the same pattern as GPT-3 again, building bigger models to achieve better results.

> There are times where a AAAA record points to an IP that doesn't respond, but the A record responds and so on > If it went down for months I wouldn't notice.

you are not alone.

I am building a service to monitor such events. usual up-time monitors mostly checks the v4 address.

https://v6monitor.com/tools/ipv6-check

Did you at least report it to scaleway, so they can fix the issue?

SVCB DNS records

looks like so, the emojies are dead give away.

I am not against using LLMs, but the author should have validated the contents before posting.

tip about ghcr.io, you only need to add NAT to the docker endpoint, their blob endpoint have v6.

I have blog about that, if you are interested: https://blog.miyuru.lk/ipv6-docker-registries-2025/#github-c...

thanks for the post.

did you reach out to honeycomb regarding the issue?

also one other trick it to use a cf warp tunnel to get v4.

https://github.com/ViRb3/wgcf

if you want population measurements, there is a APNIC page for that too.

https://stats.labs.apnic.net/v6pop

Fair warning, this page is not optimized and takes a lot of resources to render.

crossed 50% on Mar 28, 2026, 3 weekends back.

google published the latest data only yesterday, hence the delay.

>My last chance is with Scaleway.

I have been using for sw for object storage for a while. I haven't had a a issue with them with object storage.

only problem I had was with a another product called cockpit(they ship metrics about object storage there), which they bundled with object storage product, which cannot be disabled.

This month I had a 12 euro surcharge, because they enabled something in cockpit by default a while ago and suddenly start to charging for it.