verdverm

20 ms·

601.
by verdverm 2mo ago
you can also buy GitHub Enterprise onprem and get a managed, self-hosted experience, thereby avoiding the public GitHub downtime downtime and self-hosting is not a reason most people are convinced to move forges, it takes a lot to get someo
602.
by verdverm 2mo ago
> Agency independence isn’t coming back unless the Supreme Court reverses itself. Or we do constitutional amendments, which has broad appeal (>70%, bipartisan), people across party lines see the corruption from both parties https:&#x
603.
by verdverm 2mo ago
https://archive.is/5KBEa
604.
by verdverm 2mo ago
use oauth buttons, it's a single click for users with apple/google (i.e.) there are several auth SaaS providers you could look to for inspiration or adoption tl;dr - clicking a few buttons at prompts is much less effort (and more
605.
by verdverm 2mo ago
no thanks, not interested ;] atproto is a red flag for me now, toxic leaders at bluesky, who effectively control the protocol, it's not really open for others to have a say in what it looks like you also don't have private repos a
606.
by verdverm 2mo ago
> We have been under heavy load the past day or so, we're working round the clock to scale! GitHub is saying the same thing re: pushing, when I looked many months ago, it looked very much like the rust/nix/jj club, who are
607.
by verdverm 2mo ago
> you can self host your own AppView so it's never down for you Self hosted servers go down too and why people have picked hosted solutions like GitHub, Bitbucket, and GitLab
608.
by verdverm 2mo ago
Why would I want my Ai using a web page instead of an API or CLI? Seems wasteful to spend tokens on image processing when a more deterministic, text-based alternative is/should be available
609.
by verdverm 2mo ago
Tangled is forcing some of their particular technologies (nix/rust) on users for those pieces and that is a hard sell. I don't use those techs, so I'm not interested to use a platform that I need to learn them to use. fwiw, t
610.
by verdverm 2mo ago
nit, CUE is independent from Google now, Marcel left to work on it full time years ago with some other folks, they started a company
611.
by verdverm 2mo ago
Laguna wwa standout in that they borked the quants released with the main model and had to update the next day
612.
by verdverm 2mo ago
Not the first time this has been used, despite what author wants to claim, not the first time any of this has been said or commented about on HN https://hn.algolia.com/?q=ai%3Bdr This piece is still slop, human or ai create
613.
by verdverm 2mo ago
Fwiw, Microsoft has not been as destructive to GitHub as people predicted. The complaint these days is around uptime, which is understood to be an effort to migrate to Azure. It's one thing to have low quality products, it's anoth
614.
by verdverm 2mo ago
doesn't matter, it's DOA anyway, Elon is bad for your brand, hosting your project with Elon is a sure way to have people pass on it, and possibly a way your private repos end up in an Ai training set. These people cannot be truste
615.
by verdverm 2mo ago
previously discussed today: https://news.ycombinator.com/item?id=49334209 Ironically this had an issue because of GitHub downtime? https://status.cursor.com/incidents/l9h9vrd726jv
616.
by verdverm 2mo ago
> Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that". Found in the HN commenting guidelines, linked at the b
617.
by verdverm 2mo ago
We are more likely to find out the anti-network effect of Elon There are many people who will not touch this and look down upon those that do. Ethics and morals still seem to have some weight
618.
by verdverm 2mo ago
It's ongoing and seems more than a foray at this point, the worlds richest person spending heavily on politicians. Thank you SCOTUS for making unlimited money in politics legal, you really united the citizens with that one
619.
by verdverm 2mo ago
you'd be crazy to not be taking snapshots on the regular, many good reasons besides failures
620.
by verdverm 2mo ago
brand is their power, they'd be wise to not wreck it with dumb moves or PR statements (they already have some)
621.
by verdverm 2mo ago
we are in the process of transitioning from hype to commodity with llm tokens, moats are typically at the top of the stack or in the data warehouse
622.
by verdverm 2mo ago
I don't understand why they don't look for large substring matches for the system prompt before returning the response. Trivial calculation compared to a system prompt instruction asking the model not to do it
623.
by verdverm 2mo ago
Fireworks or OpenCode Go
624.
by verdverm 2mo ago
For all the talk about over reasoning, K3 on low thinking has been rather nice
625.
by verdverm 2mo ago
quanting is actually cheap and you can compress a model that does not fit on a GPU. You can process layer by layer, this is what the sequential processor in llm-compressor does.
626.
by verdverm 2mo ago
Ai responses are against HN commenting rules, lease review the guidelines linked at the bottom of most pages
627.
by verdverm 3mo ago
that control (so I can turn it off completely) is why I picked DDG to replace Google, who force feeds us the hallucinations I've stopped using DDG now because of result quality. I now use a "meta" search backed by EXA, Tavily
628.
by verdverm 3mo ago
> AI is not yet generating high fidelity solutions full stop, Ai is not as good as you believe it to be, look beyond the curtain It's not a side, it's reality, unless you believe there are multiple realities, which is plausible
629.
by verdverm 3mo ago
I would not be surprised if there are actually US components in all the military stuff too. Ukraine continues to find them in Russian weapons, surely they are in Chinese and Iranian munitions too.
630.
by verdverm 3mo ago
Should Microsoft not provide such a person? Surely they have the money and at least one person who would be acceptable to the curl team as an outside collaborator

More ›