sebmellen

8 ms·

91.
by sebmellen 3mo ago
That job description feels so far beyond parody that I could scarcely believe it until opening the link! What a world.
92.
by sebmellen 3mo ago
Are you using a local model or one of the hosted ones? I don't think a harness can influence the range of outcomes that much.
93.
by sebmellen 3mo ago
Wow, I like that a lot more than what it gave me! Shows you how wide the range of outcomes can be.
94.
Show HN: Git-temp (repo scratchpad for AI agents that won't clutter Git status)(github.com)
2 points by sebmellen 3mo ago|0 comments
95.
by sebmellen 3mo ago
Yagil, thanks for the credits! Lots of fun to try running this locally. I don't understand the negativity in some of the other comments. This is one of the better agent harnesses I've seen for inspecting reasoning chains, which is
96.
by sebmellen 3mo ago
Very neat, this is definitely a more complete prompt than mine!
97.
Kimi Code(kimi.com)
2 points by sebmellen 3mo ago|0 comments
98.
by sebmellen 3mo ago
My testing prompt for these models is by no means objective or repeatable (like the pelican) but it's a nice test of curiosity: > Impress me with a 1 page html file Result: https://ydaurtg3fdwhq.kimi.page/ Came ou
99.
by sebmellen 3mo ago
> the figure is before contractor payouts Ok… so it’s GMV, not ARR… Can we say eBay hit a gross run rate of $80B?
100.
by sebmellen 3mo ago
Yeah, but as you say, this is another LLM rewriting it. The amount of noncontextual information is nauseating and destroys the point of a README (in my humble opinion). A human might have written a disclaimer like this: > When not using
101.
by sebmellen 3mo ago
It’s so painful to read the LLM-compressed explanations. I can’t exactly identify what it is, but it’s an immediate tell and literally requires twice the effort to comprehend. For example: > Honest caveat, visible in the clip: the pxpip
102.
by sebmellen 4mo ago
I’ve found unexpected success in using ephemeral NixOS VMs for local development… once you authenticate your agent you can let it run wild without worrying about permissions.
103.
Show HN: Git-temp – scratchpad folder for AI agents; doesn't clutter Git status(github.com)
4 points by sebmellen 4mo ago|0 comments
104.
by sebmellen 4mo ago
This is really far too broad a brush. Do most medical students publish useless case studies trying to jockey for residency spots and signal hustle/devotion? No doubt! But there are a good handful of medical students who are still (surp
105.
Show HN: Git-temp (local scratchpad for AI agents that won't clutter Git status)(github.com)
2 points by sebmellen 4mo ago|0 comments
106.
by sebmellen 4mo ago
I suggest we send this fellow to the Monty Python Argument Clinic https://youtu.be/TpQlyUjp3vM .
107.
by sebmellen 4mo ago
Very cool concept with AltTab. Have always been looking for something like it!
108.
by sebmellen 4mo ago
Generally speaking, every YC company post ~2020 is forced to make pathologically false claims to compete in the (fundraising) market.
109.
by sebmellen 4mo ago
Agreed
110.
by sebmellen 4mo ago
Just commenting for posterity… if this is what it claims to be, I am not looking forward to how it will empower the people who submit bug bounties to us. Historically they’ve been people from certain identifiable countries (usually developi
111.
by sebmellen 4mo ago
Ok. Claude will not work for this use case because none of the sample data (weirdly blurry ID images) is in the training data.
112.
by sebmellen 4mo ago
Great, let me know when those models can run on-server and process/analyze streams of ID images with less than 100ms of latency. You’ll need to make sure you have a massive set of training data including all manner of slightly blurred
113.
by sebmellen 5mo ago
I would never disgrace you with reading my LLM output unless I explicitly identified it
114.
by sebmellen 5mo ago
If I'm interpreting this correctly, GitHub will use their existing actions infrastructure to run versions of the code in isolated worktrees. I think this could be a very beneficial process. What I've done on my end is created a pr
115.
by sebmellen 5mo ago
The most frustrating thing to me is to receive a 5-paragraph-plus email that was clearly written with some AI that filled in the email with vapid and useless talking points, like "Let me know if you need any other blah blah blah; While
116.
by sebmellen 5mo ago
This article is amazing because it talks at length about a magnificent video that is never shown.
117.
by sebmellen 5mo ago
Quantum itself is the most scummy, grift-filled industry. Every quantum company is riding the AI/semiconductor hype wave with basically zero revenue prospects or long-term application of the tech. Companies trading at 200x earnings, IO
118.
by sebmellen 5mo ago
In my experience, the biggest benefit comes from having good quality integration and unit tests that are easy for the agent to run on its own to verify its work against.
119.
by sebmellen 5mo ago
Same here. You’d think they could at least separate out the GitHub-hosted and self-hosted runners, so you’re still able to dispatch jobs if the self-hosted runners are down.
120.
by sebmellen 5mo ago
> Just realized that I have done all my work in @superset_sh since Dec 26. Just FYI the first quote on your site references a date we haven’t reached yet!

More ›