HackerLangs
TopNewTrendsCommentsPastAskShowJobs

postalcoder

3,037 karmajoined 2 ปีที่แล้ว
developing llm systems

side: https://hcker.news

ryan [at] hcker [dot] news

Submissions

Screens in Screens in Screens

screenpond.cool
3 points·by postalcoder·7 เดือนที่ผ่านมา·0 comments

Show HN: hcker.news – an ergonomic, timeline-based Hacker News front page

hcker.news
180 points·by postalcoder·ปีที่แล้ว·82 comments

comments

postalcoder
·เมื่อวาน·discuss
The only way to understand those moments where you sometimes Claude acts mysteriously is to be terminally online.

https://www.google.com/search?q=claude+usage+draining+site:w...
postalcoder
·เมื่อวาน·discuss
Going to repost my comment from below:

Codex creates a new folder in `~/Documents` in iCloud drive/OneDrive for every single thread you make. Furthermore, these threads are also polluted with your global AGENTS.md file, as well as all other developer_instructions that are injected by the harness.

I want my chats isolated from work contexts.
postalcoder
·เมื่อวาน·discuss
Same. I don't see why they thought this was a good move.
postalcoder
·เมื่อวาน·discuss
> Even though it says Work or Codex or whatever you can just ask it whatever you want, and it will work.

No it's not. Codex creates a new folder in `~/Documents` in iCloud drive/OneDrive for every single thread you make. Furthermore, these threads are also polluted with your global AGENTS.md file, as well as all other developer_instructions that are injected by the harness.

I want my chat isolated from work contexts.
postalcoder
·เมื่อวานซืน·discuss
I just installed this. I am very confused. I no longer have a Codex app on my computer. ChatGPT is now Codex.

But what happened to ChatGPT? Where am I supposed to casually chat?

Also, when you toggle btween ChatGPT Work and ChatGPT Codex, nothing changes. This is super confusing. Can someone from the OpenAI team clarify the difference btwn the modes? Does chatgpt work have more business-y related plugins turned on by default?

Edit: So it seems like the only place you can actually chat with chatgpt is in an awkward homeless nested window. idk. The chatgpt interface wasn't great (desperately needed artifacts), but I still used it a lot. I can't see this change going well with a lot of the casual users.

Edit2: In their awkward homeless nested chat mode, you cannot even edit past messages. this is a mess, why was the team so zealous to pull the switch on unification in this state? guessing there was internal pressure to juice codex's growth but, based on what im seeing, they did it by torching chatgpt?

Edit 3: Ok so it seems like ChatGPT is still around, but renamed to "ChatGPT Classic". Seems like it wont be long for this world because there's no place to download ChatGPT Classic should you choose to uninstall it. The dmg at https://chatgpt.com/download/ only contains the new ChatGPT.
postalcoder
·เมื่อวานซืน·discuss
> Avoid generic brevity instructions: GPT-5.6 is more sensitive than GPT-5.5 to instructions such as “Be concise,” “Keep it short,” or “Use minimal text.”

RIP Caveman skill. Six month good. Now skill dead.
postalcoder
·เมื่อวานซืน·discuss
It looks like ChatGPT finally has their own form of artifacts, which is quite welcome. I can't believe it's taken them two years to copy Anthropic on this. Artifacts were a huge differentiator btwn claude and chatgpt. I wonder why it took them so long?
postalcoder
·เมื่อวานซืน·discuss
It looks like ChatGPT FINALLY has artifacts. I can't believe this took over a year or two, it was easily Claude's biggest edge.
postalcoder
·เมื่อวานซืน·discuss
There is so much less drama involved with the Codex world. You don't realize how oppressive CC is until you've escaped it. Outages, weird restrictions, degradation, accelerated usage, etc etc etc.
postalcoder
·เมื่อวานซืน·discuss
I wish there was a comments cooldown that paused all comments until we've all been able to use this model for a few hours.

This thread is going to be thousands of comments long about a product we haven't gotten to use yet.
postalcoder
·เมื่อวานซืน·discuss
cant wait to try this with frontend. gpt-5.5 litters pages with microcopy everywhere.
postalcoder
·เมื่อวานซืน·discuss
Not their first try. There’s been reporting about how they’ve kept pushing their model releases back because of underwhelming performance.
postalcoder
·เมื่อวานซืน·discuss
Fun game.

To answer #1, I'd like a 'shuffle' button that re-shuffles the order of the letters.

I'd also appreciate the ability to "bank" the surplus time from the prior words.
postalcoder
·5 วันที่ผ่านมา·discuss
I wonder if it's related that that OpenAI has found a way to cut inference costs by half, according to The Information.

https://www.theinformation.com/newsletters/ai-agenda/openai-...
postalcoder
·6 วันที่ผ่านมา·discuss
You still have to worry about misconfigured local models. Even the professionals get it wrong, which is why local model performance is uneven across providers.
postalcoder
·8 วันที่ผ่านมา·discuss
I did find the bit about ChatGPT's crappy prose amusing because while that may feel like it's a good benchmark for the state of AI, the quality of prose doesn't necessarily correlate with the major progress made over the last few years which is in post-training.
postalcoder
·8 วันที่ผ่านมา·discuss
Ok I'll take the bait.

> Look: I work at an AI company. I use AI all day.

Looks like OP works at dropbox. Dropbox is not an AI company. It's not remotely one.

Why does this matter? Because it undermines the entire point of the post. Later:

> And if you’re watching that kind of hyped content: You can be part of the solution, too. Hold your favorite creators accountable! Ask them to show you the receipts! If you know that something doesn’t work, don’t just let it slide.

AI confidence theater for me, but not for thee.
postalcoder
·9 วันที่ผ่านมา·discuss
> I feel like the problem is more easily solved by adding a chronological view or filtering to an existing tree-based discussion

I built this (both chronological view and new comment filtering) into the comments presentation on https://hcker.news. Check it out, I’d be interested to know if there’s any way I can make it more useful.
postalcoder
·9 วันที่ผ่านมา·discuss
The downside with reddit-/hn-style comment is that, while they provide a superior UI for discussions, the liveliness of the discussions have a shelf life of a day. It makes it's hard to get a high quality discussion about new/breaking topics.

What I mean is that, for new products, the threads that get the greatest discussion liquidity are those where not a single person knows a thing about it. So you'll get hundreds to thousands of comments that don't have a clue. In this world, influence concentrates around people with pre-release access to these products.

In the HN/Reddit paradigm, how do people impart their experiences with a model like Fable? You could submit a new blog post and some people will comment on that to discuss their experiences. You could do an Ask HN but those don't get much traction.

Old style forums were a pain in the butt to read but they were better for focused discussion over time.
postalcoder
·10 วันที่ผ่านมา·discuss
I made a filter that shows github repos posted to hacker news that do not have evidence of AI authorship:

https://hcker.news/?ai=exclude&include_domains=github.com