Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
nateb2022
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
61.
▲
by
nateb2022
2mo ago
browser-use is a lot more capable and generally more efficient, and a lot easier to embed in other programs than Claude's driver. I haven't used Vercel's agent browser enough to speak to it.
62.
▲
by
nateb2022
2mo ago
https://github.com/browser-use/browser-use has been great so far
63.
▲
by
nateb2022
2mo ago
official site seems to be https://deepseek.com/harness/en/
64.
▲
by
nateb2022
2mo ago
Which quant would cut it closest on a 16GB VRAM card?
65.
▲
by
nateb2022
2mo ago
Currently using https://github.com/placemark/check-geojson for this usecase right now. Noticed your code includes a custom JSON parser, which seems to silently overwrite duplicate keys, which is absolutely something a
66.
▲
by
nateb2022
2mo ago
The bench is not a black box, if you scroll down it gives an example of what the test cases look for
67.
▲
by
nateb2022
2mo ago
[dupe] https://news.ycombinator.com/item?id=49299605 (3 hr ago, 217 comments)
68.
▲
by
nateb2022
2mo ago
On the contrary; it seems Luna (max) is 2x as capable at 1/5 the cost: https://cognition.com/frontiercode
69.
▲
by
nateb2022
2mo ago
This was the older Gemini 3.7 submission and linked to the blog post; the submission you marked this as a dupe of occurred 20 min after this one, and linked to the much more barren https://ai.google.dev/gemini-api/docs&
70.
▲
by
nateb2022
2mo ago
Or a 3.7 Flash-Lite
71.
▲
by
nateb2022
2mo ago
[dupe] https://news.ycombinator.com/item?id=49288847 (35 points, 8 comments)
72.
▲
by
nateb2022
2mo ago
Codex CLI should have been usable for a while now, and you could also sign in to your ChatGPT subscription using Opencode which runs perfectly fine on a Raspberry Pi.
73.
▲
by
nateb2022
2mo ago
That's my bet, I doubt they'd have the resources to have trained anything substantially larger
74.
▲
by
nateb2022
2mo ago
Yup. I posted about the huggingface countdown: https://news.ycombinator.com/item?id=49272534 but seems like both the huggingface page, and the modelscope.cn URL that commenters found, were quickly taken down. Not sure if th
75.
▲
by
nateb2022
2mo ago
This modelscope.cn url, like my huggingface url, seems to have been taken down.
76.
▲
by
nateb2022
2mo ago
URL was taken down; seems like the modelscope.cn link that other commenters posted also got taken down.
77.
▲
by
nateb2022
2mo ago
It used to feature a countdown, at least when I posted. 2 days and 2 hours from Wednesday 12 August at ~11am. Huggingface or Qwen took down the URL it seems. Edit: looks like the modelscope.cn URL that other commenters posted also got taken
78.
▲
Qwen 3.8-27B goes openweight in 2 days
(huggingface.co)
19 points
by
nateb2022
2mo ago
|
22 comments
79.
▲
Anydoc: Word, PowerPoint, Excel, OpenDocument, RTF, ePub, CSV, PDF to Markdown
(github.com)
3 points
by
nateb2022
2mo ago
|
0 comments
80.
▲
by
nateb2022
2mo ago
Think of an LLM as a thesaurus, but for entire trains of thought rather than words. Your initial query yields something pertinent to the task at hand. But let it endlessly recurse and... you end up with something completely useless. People
81.
▲
Ling 3.0 Flash: openweight 124B-A5.1B competitive with MiniMax M2.7
(huggingface.co)
3 points
by
nateb2022
2mo ago
|
0 comments
82.
▲
LFM2.5 2.6B model competitive with 4x larger models
(huggingface.co)
169 points
by
nateb2022
2mo ago
|
40 comments
83.
▲
Macro: Unified email, messages, docs, tasks, agents, CRM for teams
(github.com)
5 points
by
nateb2022
2mo ago
|
0 comments
84.
▲
by
nateb2022
2mo ago
Seems to be entirely AI written. Full of the stereotypical rule of 3, and "not X, it's Y," in fact I think each paragraph features at least an instance of one of them if not both.
85.
▲
by
nateb2022
2mo ago
Some well written devblogs are also available: - https://lwss.github.io/Duty-Of-Kisak/ - https://lwss.github.io/Kisak-Black/
86.
▲
DCSystems: Independent tech talk series in the DMV area
(dcsystems.xyz)
2 points
by
nateb2022
3mo ago
|
0 comments
87.
▲
IBM Claims Quantum Advantage with New Validation Techniques
(spectrum.ieee.org)
7 points
by
nateb2022
3mo ago
|
0 comments
88.
▲
by
nateb2022
3mo ago
I use Luna a lot (over 1T tokens since it came out) and I'd rank Luna (high/xhigh) on par with Sonnet 5, without hesitation.
89.
▲
Openinterpreter: Terminal coding agent built for low-cost models
(openinterpreter.com)
2 points
by
nateb2022
3mo ago
|
0 comments
90.
▲
Audio8 TTS Preview 0.6B: multilingual TTS model with zero-shot voice cloning
(huggingface.co)
2 points
by
nateb2022
3mo ago
|
0 comments
More ›