logankeenan

6 ms·

31.
by logankeenan 11mo ago
This is great! I had to create an xhr-fetch-proxy to use fetch and htmx. This change should open up a lot of fun possibilities. https://github.com/logankeenan/xhr-fetch-proxy
32.
by logankeenan 11mo ago
> Usage their customers would not pay for if it did not have a positive ROI. I don't think we can assume that's true. Their customers are paying for it, but we don't know how profitable they are being with the AI compute t
33.
by logankeenan 11mo ago
Can you elaborate on this a bit more? Does that mean OpenAI had a ~$9.4 billion loss so MSFT needs to put 49% of that loss in their books?
34.
by logankeenan 1y ago
I'm on the binary install with version v2.0.19. It never showed in the `/model` selector UI. I did end up typing `/model haiku` and now it shows as a custom model in the `/model` selector. It shows claude-haiku-4-5-20251
35.
by logankeenan 1y ago
I'm not seeing it as a model option in Claude Code for my Pro plan. Perhaps, it'll roll out eventually? Anyone else seeing it with the same plan?
36.
by logankeenan 1y ago
I’d really like to understand the total compute cost it takes to accomplish these tasks. I assume the compute is happening in a DC somewhere and not all onboard. Is the total cost of compute plus electricity to power the machine less than t
37.
by logankeenan 1y ago
It’s great that they made an update. It’s not clear to me if this can result in a RCE. If it does, then does this mean that enough iPhone 6s are still out in the wild where a bad actor could easily take over a big enough portion to do more
38.
by logankeenan 1y ago
Has anyone had luck with bat boxes to help control the mosquito population?
39.
by logankeenan 1y ago
Iran is the largest state sponsor of terrorist groups. The key word being “state”. There are many well known terrorist groups that are not sponsored by Iran.
40.
by logankeenan 1y ago
With Electron, it will bundle chrome into the app so you only have to handle that single rendering engine. Tauri uses whatever is the default browser on the system the app is installed on
41.
by logankeenan 1y ago
Do you mean the actual rendering of html/css when you say rendering differences? Or are you referring more to differences in JS support?
42.
by logankeenan 1y ago
Do you have a source for that? Recently, I was trying to understand the trade deficit when consider services, but couldn’t find anything credible
43.
by logankeenan 1y ago
I used molmo to parse screenshots in order to detect locations of UI elements. See the repo below. I think Omni parser from Microsoft would also work well. https://github.com/logankeenan/george https://githu
44.
by logankeenan 1y ago
I’ve been wanting to try Claude Code. What makes it such a difference maker compared to existing AI tools?
45.
by logankeenan 1y ago
That looks pretty nice and polished! Did you use Claude Code? Also, what are some of the crates you used for your Rust stack? Did you ever consider doing a server side rendered app and no React?
46.
by logankeenan 1y ago
I believe they are referring to marginal costs. Yes, with economies of scale can become cheaper to produce, but they still require capital to reach that scale. https://en.m.wikipedia.org/wiki/Marginal_cost
47.
by logankeenan 2y ago
> lets you stop breaking down one level up from what you're used to. Can you provide an example of what you mean by this? I provide very verbose prompts where I know what needs to be done and just let AI “do” the work. I’m curious h
48.
by logankeenan 2y ago
Iroh has a local discovery option which uses mDNS-like system for discovery when on the same network. https://www.iroh.computer/docs/concepts/discovery#local-disc...
49.
by logankeenan 2y ago
I’ve been experimenting with Iroh over the past month. It’s been such a pleasant experience. Previously, I created a p2p connection with webrtc and had to create a layer on top of that to use request/response style communication betwee
50.
by logankeenan 2y ago
I live in one of the states mentioned in the article. It’s been really cold here for the last two days, but being from the Midwest, you kind of get used to it. Tomorrow’s just going to be another cold day
51.
by logankeenan 2y ago
They make money from their core income streams like ads or hosting for AWS. People pay them money for these services because doing so helps those companies make money. What about AI, do they expect to sell that much AI compute to other comp
52.
by logankeenan 2y ago
I’m genuinely curious how these companies are going to get a ROI?
53.
by logankeenan 2y ago
I don’t think this answers the parent question. Database queries still needs to be made and a view still needs to be rendered. What complexity does next.js solve?
54.
by logankeenan 2y ago
If you received one dollar per second from the day you were born then you’d be a millionaire in ~11 days and a billionaire in ~31 years. At some point, these big numbers are generalized as “big numbers” and we don’t conceptually comprehend
55.
Ask HN: What's the right way to do layoffs?
4 points by logankeenan 2y ago|4 comments
56.
by logankeenan 2y ago
Have you been to the localLlama subreddit? It’s a great resource for running models locally. It’s what got me started. https://www.reddit.com/r/LocalLLaMA/
57.
by logankeenan 2y ago
Do you have any recommendations on models to try?
58.
by logankeenan 2y ago
Are the upload events specifically used for files or for most server interactions. I haven’t noticed any issues with my proxy yet…
59.
by logankeenan 2y ago
I really wish htmx would use fetch rather than xhr, and now it looks like that won’t happen. Fetch is easier to proxy, so it would open up a lot more possibilities for backends, like ones running in browser or a native process. I had to hac
60.
by logankeenan 2y ago
Molmo released recently and is able to provide point coordinates for objects in images. I’ve been testing it out recently and am currently building an automation tool that allows users to more easily control a computer. Looks like Anthropic

More ›