I put them all in containers at a minimum now, for this reason. I have one click / command to export file diffs back to the host, and got reload in from the host if I want it, so there is little value to letting agents be on the host anymore

I think the community is in a phase right now, driven by the massive increase in clawd spam. So the flag button is likely getting clicked more and that has leakage.

Flagging is something the community does, not the two official moderators. What do you see changing in the FAQ?

That makes sense and is certainly legitimate. But I was noticing some politicla topics seem to be most likely to be flagged. I also sometimes see them get flagged before the thread even gets long. So I was wondering if some political topics seem to be off-limits.

Yea, the political side is another aspect, similar to Trump 1.0 if you were around HN for that, but also amped up like everything around Trump 2.0

I was not around for 1.0; it sounds colorful. The HN guidance on curiosity, kindness, being thoughtful and substansive...all solid. The same guidelines describe "politics" as "off topic" but not prohibited. I had thought there may be specific subtopics which are actually prohibited. What I probably have been seeing is that clawdbot is more active on divisive topics, because of key words, and thus gets flagged.

Generally politics gets to the front when it overlaps with some other topic HN is passionate about. Trump must make headlines each day, so the frequency of stories that overlap are elevated.

Yea, this is going to be real unhealthy for society. We have time tested methods for dealing with grief as individuals and as groups, many millennia of experience and wisdom. This kind of thing is going against all that we know about these things, letting people stay in the early stages, and even something beyond that takes them further away from that process.

Great HN handle for the theme of your submissions. Are you collecting them on a webpage or in git repo?

One of the authors is answering questions (submission is a link to that post)

There is a DRM free link provided by the author in said thread too

Ah, so this is what it is. I'm sure they will have no editorial biases in deciding what blocked content should be set free

I'll be curious to see the venn with X and Trump Social

You don't need to copy-paste the first section as a comment, it's generally frowned upon

I share this sentiments and complaints about NextJS and Vercel. I would add they prioritize designing everything for their edge paradigms, so if you don't use their products or preferred architecture, it becomes that much harder to do things.

I'm on TanStack now, very happy with the decision after more than a year.

For my custom agentic coding setup, I use Claude Code derived prompts with Gemini models, primarily flash. It's night and day compared to Google's own agentic products, which are all really bad.

The models are all close enough on the benchmarks and I think people are attributing too much difference in the agentic space to the model itself. I strongly believe the difference is in all the other stuff, which is why Antropic is far ahead of the competition. They have done great work with Claude Code, Cowork, and their knowledge share through docs & blog, bar none on this last point imo.

ChatGPT 4.5 was never released to the public, but it is widely believed to be the foundation the 5.x series is built on.

Wonder how GP feels about the minor bumps for other model providers?

Minor version bumps are good and I want model providers to communicate changes. The issue I am having is that Gemini "preview" class models have different deprecation timelines and rate limits, making them impossible to rely on for professional use cases. That's why I'd prefer they finish the 3.0 role out prior to putting resources into deploying a second "preview" class model.

For a stable deployment, Google needs a sufficient amount of hardware to guarantee inference and having two Pro models running makes that even more challenging: https://ai.google.dev/gemini-api/docs/models