> Wasn’t the original intention behind facebook to accumulate a directory of hotties
Maybe so, but have you seen Zuck's wife? I'm pretty sure he could find someone hotter to date if he cared to. There must be armies of gold-diggers after him. And yet he seems happy with his imo rather plain looking wife. Well done them both!
It's respectful to trust someone who says they're not suicidal. Saying "they could've been suicidal anyway" is disrespectful to people who aren't suicidal and are telling the truth.
If someone is struggling with suicidal thoughts and is publicly lying about it, they shall not have my respect anyway: I'm ok with being disrespectful to them.
Opus is about 7 times more expensive than GLM with API pricing. And since you can only use the Opus subscription plan in CC, you're essentially locked into API pricing for Pi and any other harness.
So you're either paying $1000's for Opus in Pi, or $30/month for GLM in Pi. If the results are mostly equivalent that's an easy choice for most of us.
Perhaps I'm being extremely daft: If the API is 7 times more expensive, then why is it $1000 vs $30? Or is there a GLM subscription one can use with Pi? Certainly not available in my (arguably outdated) Pi.
I'm not the OP, but it's the latter. I'm currently using the "Lite" GLM subscription with OpenCode, for example. I'm not using it very heavily, but I haven't come close to hitting the limits, whereas I burned through my weekly limits with Claude very regularly.
I am using GLM-5.1 in pi.dev through Ollama Cloud. I am able to get by on the $20 plan. I use it a lot and the reset is hourly for sessions and weekly overall. This is the first week I got close to the limit before reset at about 85% used. I am probably using it about 4 hours a day on average 6 or 7 days per week.
Or tell pi to add support for the coding plan directly. That gave me GLM-5.1 support in no time along with support for showing the remaining quota, etc, too.
It also compresses the context at around 100k tokens.
If the question is why Chinese models are contributing to open source and sharing of information, I don’t pretend to know the rationale but I think it’s because it’s an economic war.
I think the Chinese models have to be more open to increase trust as everyone is worried they are feeding their very essence/soul into a Chinese copying machine.
Also China wants there to be viable competitors so that US can’t just dominate a potentially very important field. It’s a challenge to a unipolar USA dominated world.
Also it helps to spur Chinese companies in the all important microchip industry which is controlled by a very small number of companies at various steps in the supply chain.
I wonder too if it allows them to hold an ace in their hand as well in terms of threat/power for negotiations. As in, they can cause the whole house of cards to crumble, an economic nuclear weapon so to speak.
Finally, there is a certain amount of prestige involved too. China can compete or even win at a very complicated game. They use it to increase national pride and to project their advancing power status to other nations.
Anyways, just my thoughts. Interested in others thoughts.
> The more involved a father is with their baby's care, the deeper this transition becomes
My partner died when my child was a year and a half, so I'm more involved than basically any father, to the point that my experience is much closer to a mother's, a role in which my only solace is seeing myself struggle slightly less than many mothers.
It's reassuring to read an article confirming I'm screwed.
Even (especially) in hindsight it will have been the greatest thing you did with your life. At least this has become clear to me (now 27 years a father).
Yeah, but we're a small company and sometimes cut corners to move faster, so if a tool can solve this instead of potentially adding more friction to other engineers I'm all for it.
> Stage automatically analyzes the diff, clusters related changes, and generates chapters.
Isn't that what commits are for? I see no reason for adding this as an after-thought. If the committers (whether human or LLM) are well-behaved, this info is already available in the PR.
In our experience, it's difficult to create well-mannered commits as you code and new ideas pop into your head or you iterate on different designs (even for LLMs). One concept we toyed around with was telling an LLM to re-do a branch using "perfect commits" right before putting up a PR. But even then you might discover new edge cases and have to tack them on as additional commits.
We thought git wasn't the right level of abstraction and decided to tackle things at the PR level instead. Curious to hear your experiences!
> In our experience, it's difficult to create well-mannered commits
Sure, it is. But it's worth it, not just for code review, but for a myriad other things: bisect, blame, log, etc.
Your tool makes one thing (the code review) easier, while decreasing people's motivation to make well-mannered commits, thus making everything else (bisect etc) worse.
I'm sure it's net positive in some cases, and I think it's net negative in other cases.
> We thought git wasn't the right level of abstraction and decided to tackle things at the PR level instead. Curious to hear your experiences!
The frick is a PR abstraction? Is this a GitHub PR abstraction where the commits are squashed and the PR description is whatever was hallucinated at 5 am? Yes, that’s certainly an abstraction, aka loss of information.
You either have the information stored in the version control database or you don’t. You can curate and digest information but once it’s lost it’s lost.
People layering stuff on top of Git or Subversion makes no sense. Your AI is not so dainty and weak that it cannot write a commit message. And if it can’t then you can recuperate the information that you trashed.
I concur. I cannot accept that we are so disconnected from what we're building that we can't go back and revise our commits or something else to make it make sense.
My understanding is that ATProto itself is definitely decentralized but the app view most people interact with using the Bluesky app is centralized ...sort of. The Bluesky app view will read from PDSes hosted by other people, hence people on Bluesky can see stuff posted elsewhere, like users of Blacksky. If the Bluesky app view decides to stop reading from any other PDS (like those of Blacksky, or ones which are self-hosted) they're free to do so. The same is true for alternative app views like Blacksky. Since most people think of Bluesky as the thing you see on the official Bluesky app (which shows the Bluesky app view) an outage of the Bluesky app view will mean they lose the ability to view any posts from any source. If someone's using a separate app view like Blacksky, the most that will happen to them should be that they'll lose interaction with posts coming from Bluesky's PDSes until the outage ends.
I may have the division between Bluesky and Blacksky off, but ATProto does allow this sort of thing. Hosting a PDS is trivial and requires very few resources. Hosting a full app view can be expensive depending on how many PDSes you're ingesting from, but you can decide how much of that you want to do.
It sounds like Blacksky's outage was much more limited and was caused by them missing some spots where their code still accidentally had some Bluesky integration due to it being the default: https://bsky.app/profile/rude1.blacksky.team/post/3mjnf6pubr...
Bluesky has never been distributed/decentralised. It's a single central system, which fetches 0.001% of user data from external systems if the user opts in, and has a marketing team that calls this decentralisation.
The Bluesky app view is centralized in that it can decide which content to show, but A) the hosting of that content is decentralized, and B) alternate app views like Blacksky exist which are fully independent of Bluesky (both Bluesky the company and Bluesky the app view). The Bluesky app view could stop showing users content from Blacksky (or any other) PDSes, but that's it. If you're using the Blacksky app view, afaik Bluesky the company can't do anything other than cut you off from Bluesky's PDSes.
> If by "decentralised" you mean "0.001% of it is not only hosted centrally"
Sure, much like how email is decentralized in theory but barely is in practice. This doesn’t mean that the decentralized nature is just a marketing gimmick.
It’s unsurprising that almost everyone uses the Bluesky app given that A) the infrastructure for hosting your own relay or app view (I can’t remember which) didn’t have a reference implementation until a while after launch, and B) the user base is much less tech-y than what I’ve seen on Mastodon. Most of the user base moved over in the flight from Twitter/X a couple years ago. I think if it had come out at a different time you’d see something which looked a lot more like Mastodon’s large population distribution.
Also, while this doesn't really matter it looks like the number of users on non-Bluesky PDSes is 1.42% of the total, not 0.001%.
> They have designed a protocol that could theoretically be decentralised. Then reality hit, and it was centralised.
Could you explain what you mean by the underlying protocol having become centralized over time? While I can understand arguing about whether or not Bluesky-the-social-network is practically decentralized to the degree of something like Mastodon or that it became more centralized over time, I think arguing that ATproto[1] itself isn’t decentralized would be ludicrous.
> Sure, much like how email is decentralized in theory but barely is in practice.
Hard disagree. Email is very much decentralized. Doesn't mean that there's still a long tail distrubtion, but its not like 99.999% of email accounts are on Gmail. And I can set up an email account in a few minutes and by choosing from a list of thousands of providers all over the world.
It looks like about 97% of people use iCloud, Gmail, or Outlook as their provider. That doesn't feel like it's terribly different. The people not using one of those big three make up a comparable percent of the total as the number of Bluesky users using alternative PDSes.
Three providers vs one provider does feel very different if those are the only option, but that's not the case for email or Bluesky/AT proto.
> I can set up an email account in a few minutes and by choosing from a list of thousands of providers all over the world.
You can also set up an account on a different PDS and/or use a different app view quite trivially too so I'm not sure that's substantially different.
ATProto would need to use signing key cryptography and content addressable storage to be distributed. If we can't store our data with third parties or create an offline-first system then it's not a decentralized social network.
ATproto does support storing data elsewhere. That’s what a PDS does. I’m not sure what you mean by an offline-first system in this context though or why it’s required for decentralization. Could you elaborate?
reply