user34283

10 ms·

31.
by user34283 5mo ago
The designing and implementing of a code harness in your workflow can be as simple as running something like /skill-builder. You prompt for what you want it to do, and it will write eg. python scripts as needed for the looping part, an
32.
by user34283 5mo ago
I use the Codex desktop app. In the GUI I can see the context indicator and usage stats. It also makes it easier to jump between conversations and see the updates. Sometimes I use Claude Code or opencode in the terminal, and my experience i
33.
by user34283 5mo ago
Velocity without direction? I‘m personally directing and QA testing every feature. I don’t know how socioeconomic concerns, product strategy, and mental health are a concern for me here. I‘m having a great time with my project and it’s been
34.
by user34283 5mo ago
It’s in-app on iOS. Ironically, already another user in this comment section was concerned about the security of my nonexistent backend. But it’s good to know, I was not previously aware that video processing on the backend is a common sour
35.
by user34283 5mo ago
I would be impressed if you implement it in an hour with the following features: - webview fallback with canvas capture for codecs not supported in the default player - detecting blank frames and diff between thumbnails to maximize variety
36.
by user34283 5mo ago
In my experience Codex barely writes any comments, despite my attempts to encourage it in the AGENTS.md.
37.
by user34283 5mo ago
I feel pretty good about the odds of attackers exploiting security holes in image editing functions my app does not have, in order to enter my also nonexistent datacenter or vpc.
38.
by user34283 5mo ago
On the other hand, I don’t need to review carefully every line of code in my thumbnail generator and associated UI. My nonexistent backend isn’t going to be pwned if there is a bug in the thumbnail generation. After the QA testing on my dev
39.
by user34283 5mo ago
In my opinion you are just wrong. It’s an absolute game changer, and it can now multiply your productivity fivefold if it’s a solo greenfield project. Maybe half a year ago it was as you said. You had to wait for the agent to finish, you ha
40.
by user34283 5mo ago
Other comparisons: About equal to a major iOS update at 8 GB x 1.5B. Netflix and YouTube together are perhaps around 200EB/month.
41.
by user34283 5mo ago
It’s a reasonable expectation from a clearly successful and competent engineer who is using the latest tooling. Who is to say that it’s wrong?
42.
by user34283 5mo ago
In a large codebase it‘s probably next to impossible to get people who fully understand the code to explain it to you with unerring accuracy. AI can get a pretty good picture, near instantly, whenever you need it. It’s not just competent-so
43.
by user34283 5mo ago
And there’s some truth to it. I expect tomorrow’s models will be so much more capable that we will happily pay more. But if not, we will still likely get today’s capabilities or more for cheap. I don’t see a realistic scenario in which the
44.
by user34283 5mo ago
How do you reconcile these ideas with the fact that cheap open weight models are only slightly behind the state of the art? If anything, I would bet that next year you could get today’s flagship performance for significantly cheaper via an
45.
by user34283 5mo ago
I looked into this long ago, and imo it doesn’t look as bad as you say. Musk downplayed his correspondence and willingness to meet with Epstein to the point where you could argue Musk was lying, yes. However, he did decline an invitation to
46.
by user34283 5mo ago
Kimi K2.6, the leading open-weights model, is $0.95 input / $4.00 output / $0.16 cache. That’s about 1.45x more expensive than K2.5 from January. It is around 5x cheaper than GPT 5.5.
47.
by user34283 5mo ago
I don’t think that makes much sense, surely as a billionaire you don’t need to consort with Epstein to meet women around 25 years old. That link seems to report on the same single WSJ article that mostly alleges workplace power-balance issu
48.
by user34283 5mo ago
Luckily inference is cheap, and other providers offer efficient models. It’s only going to get better in the future.
49.
by user34283 5mo ago
I never heard about the horse related thing, that’s interesting, thanks. A long history? Another search tells me that apart from the mentioned accusation, there is only one WSJ article alleging sexual conduct with SpaceX employees. You aske
50.
by user34283 5mo ago
My searches have not turned up a result showing that Musk "claimed he had never corresponded with Epstein". Can you source this? If not, can you explain why you did not check it before you posted the inaccurate claim?
51.
by user34283 5mo ago
On Artifical Analysis it shows only Kimi K2.6 and Mimo V2.5 Pro as better. Those models are 1T parameters total and 30B or 40B active, this might make abliteration impractical. About Musk, yes, there is correspondence. The only confirmed me
52.
by user34283 5mo ago
It helps if you have a monopoly on app distribution for half of all phones, or video streaming. Then you can afford zero support and still take 15-30%.
53.
by user34283 5mo ago
Does it do anything that GitLab does not? With what I heard about GitHub Actions, the GitLab CI pipelines should be much better. Not that I haven’t shot myself in the foot with GitLab pipelines on numerous occasions.
54.
by user34283 5mo ago
Hundreds or 10k+ users? I imagine requirements and integrations may differ a lot. I have seen many incidents with a large instance.
55.
by user34283 5mo ago
I am using the Codex desktop app without the pi harness and my experience is quite different. 5.5 has been a noticeable improvement over 5.4, solving more complicated issues and faster too. 5.5 does not use a huge amount of my session limit
56.
by user34283 5mo ago
To evaluate the legal risks of using AI generated code, let’s consider how many lawsuits there have been over these concerns. Inadvertent copyleft license violations: probably 0 lawsuits Competitor copied your software, you could not defend
57.
by user34283 5mo ago
We also pay $300/month for Azure Desktop VMs. We are paying for tens of thousands of those machines, although everyone knows they are stupidly expensive and incredibly slow.
58.
by user34283 5mo ago
I will see if that becomes a blocker. There was one feature/screen that Codex built in a single 5k LOC file. It was still perfectly capable of developing the feature and it was working as expected. I had it break it down into multiple
59.
by user34283 5mo ago
One, thus the git worktrees. You might think that this would lead to a mess with merge conflicts, but the agent can resolve them automatically. I added an instruction to AGENTS.md so that before handoff it fetches and rebases, resolving con
60.
by user34283 5mo ago
At this point it’s undeniable for my use cases. After I discovered how to use git worktrees in Codex to work in three conversations in parallel, I am able to build apps with a scope that simply was not realistic before.

More ›