Hacker Newsnew | past | comments | ask | show | jobs | submit | biglyburrito's commentslogin

This is also what I'm expecting. The change to usage-based billing may significantly change how much it costs to build apps.

You have to be pretty stupid to believe that "reason".

The Colorado senators (23 Democrats, 12 Republicans) that unanimously voted for the anti-right-to-repair bill (https://legiscan.com/CO/rollcall/SB090/id/1684182):

* Amabile, Judith "Judy" [D]

* Baisley, Mark [R]

* Ball, Matt [D]

* Benavidez, Adrienne [D]

* Bridges, Jeff [D]

* Bright, Scott [R]

* Carson, John [R]

* Catlin, Marc [R]

* Coleman, James Rashad [D]

* Cutter, Lisa Ann [D]

* Danielson, Jessie [D]

* Daugherty, Lindsey N. [D]

* Exum Sr., Thomas "Tony" E. [D]

* Frizell, Lisa [R]

* Gonzales, Julie [D]

* Hinrichsen, Nick [D]

* Jodeh, Iman [D]

* Kipp, Cathy [D]

* Kirkmeyer, Barbara [R]

* Kolker, Chris [D]

* Lindstedt, William [D]

* Liston, Larry G. [R]

* Marchman, Janice [D]

* Mullica, Kyle [D]

* Pelton, Byron [R]

* Pelton, Rodney "Rod" [R]

* Rich, Janice [R]

* Roberts, Dylan [D]

* Rodriguez, Robert [D]

* Simpson Jr., Cleave Alan [R]

* Snyder, Marc Alan [D]

* Sullivan, Tom [D]

* Wallace, Katie [D]

* Weissman, Michael "Mike" [D]

* Zamora Wilson, Lynda [R]


Have any of them made a statement on why they voted this way?

n.b. I'm not looking for simple "because of money/lobbying" speculation. I'm interested in the reasoning straight from the horse's mouth.


OK, so it passed the state Senate unanimously and died in the House.

The HN audience in particular might appreciate that some of the bill text revisions are dated 1969-12-31! :P


That's sickening. Every single one of these people should be primaried and removed from office.

Can't say I blame them.

My nephew graduated with a CS degree a few months ago & I have no idea if he's had any luck finding an entry-level job in his field. I'm a senior engineer w/ 25+ years experience & it's been extremely difficult to stay employed. Nor are our experiences anomalous; I know plenty of folks who are 6-24+ months out of work in tech, and not for lack of trying to find a new job.


I'm on the other side of the fence. We're just about done migrating from GitHub to GitLab (self-hosted) and it's been refreshing to DGAF about any of the GH outages I read about.

Similar boat myself too, finished moving all important stuff from GitHub to self-hosted Forgejo with cross-platform builds. Not only do I avoid all the downtime stuff, but E2E builds also takes ~20% of the completion time it used to take, since now my runners have dedicated hardware hosted at home.

To maintain a fair comparison, GitHub has supported self-hosted runners for several years (maybe that doesn’t work for your specific usage, for whatever reason).

> To maintain a fair comparison, GitHub has supported self-hosted runners for several years

Yeah, tried that first, as I didn't want to move to Forgejo, I just wanted to keep working when I wanted to work.

The GitHub runner on Linux seemed fine, but the ones for macOS and Windows seemingly did something that made them a hell lot slower than even running VMs and then executing stuff inside those. I'm not sure what the runner is doing, if there is some built-in sandboxing or what not for those platforms, but it wasn't feasible to rely on for me as the builds took way too long time.


We were on self-hosted Gitlab but after a merger were forced to Github. Navigation feels painful in comparison and basic features such as commit graph are now behind more expensive tiers.

> We were on self-hosted Gitlab but after a merger were forced to Github. Navigation feels painful in comparison and basic features such as commit graph are now behind more expensive tiers.

Same experience here. Add to that that even on Enterprise tier:

- 1 Enterprise : 1 namespace - although you can segment it with Orgs, we were advised not to do it because we're too small (~2k people) (GL: groups, subgroups, sub-subgroups, ...)

- SSH deploy keys are singletons across the entire instance and repo-bound (and Weblate for instance can only use its own key), so you need a service account for that (GL: instance-wide SSH deploy keys that you can activate in specific repos)

- GHCR only really supports classic PATs for authentication ( https://docs.github.com/en/packages/working-with-a-github-pa... - GL: proper deploy keys properly inherited throughout the hierarchy)

So all in all the experience so far is a huge step-down. I really liked pinning commonly accessed pages in the sidebar.


Interesting! I worked with Gitlab and I also thought it was quite clunky. If it was not for the stability issues GitHub is fine. Any other alternatives to GH or GL?

Self-hosting with open source code:

- SourceHut: https://sr.ht/~sircmpwn/sourcehut/

- Forgejo (used by Codeberg, etc.): https://forgejo.org/


SourceHut never really clicked for me. It doesn't give me anything useful that I don't already have in a bare git repo through a ssh.

Forgejo, on the other hand, is a drop-in replacement for GitHub.


Also:

- https://about.gitea.com/ (F/OSS MIT license self hosting GitHub like instance)


We switched from Bit bucket to Gerrit internally and it was a steep learning curve for the des but it's fine.

At a customer we're implementing GitHub Actions and even on our Dev environment there are so many hickups with GitHub.



Gitea might be an option also.

Jira / Bitbucket / Teamcity.

Might be pricy though.


Having used Teamcity for CI I cannot think of a more clunky and hard to use system (compared to GHA, which is what we migrated to).

How are you using UBO to do that?


The crudest way possible (via custom filters) - when that stops working I'll likely just do a browser plugin.

    news.ycombinator.com##tr.athing:has(a:has-text(/AI/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/AI/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/claude/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/claudeI/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/llm/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/llm/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/vibecode/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/vibecode/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/agi/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/agi/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/deep learning/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/deep learning/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/agent/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/agent/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/TPU/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/TPU/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/GPT/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/GPT/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/DeepSeek/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/DeepSeak/i)) + tr
    news.ycombinator.com##tr.athing:has(a:has-text(/Anthropic/i))
    news.ycombinator.com##tr.athing:has(a:has-text(/Anthropic/i)) + tr
It makes HN more like what I appreciated about HN in the first place by removing (what to me) is noise and increasing the signal.


Inshallah.


And honestly, what other types of questions would you ever need answers to?


I wonder how long it will take China to respond to the release of Mythos and what their response will look like.


Which works fine, right up until China releases a new DeepSeek model that's 85% as capable as an Anthropic or OpenAI premium model but costs a fraction of what either of those US companies are charging.

Speaking of which:

https://www.cnbc.com/2026/04/24/deepseek-v4-llm-preview-open...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: