mips_avatar

5 ms·

1.
by mips_avatar 4d ago
The problem with being a safety focused AI lab, is you're also a danger focused AI lab. I don't think being danger focused leads you to build inspiring things.
2.
by mips_avatar 9d ago
We really need some open source project to manage all the shims across the world. Like in Seattle the WSDOT API has live ferry locations but they never made a GTFS feed for, so i made my own shim for it. I feel like there needs to be some
3.
by mips_avatar 20d ago
As far as I can tell is these standards (MCP/MHS/etc) are just semi obvious tool inferfaces that Anthropic uses as training scenarios.
4.
by mips_avatar 21d ago
Surprised Amazon managed to fumble mechanical turk at the same time Mercor/Scale and all these other companies started hiring humans to do data labeling tasks.
5.
by mips_avatar 22d ago
I spent multiple hours every week in various risk management meetings one semester for our drone building student team. There was no risk management done, they just made it clear they would try and expell and sue us if we tested our drone.
6.
by mips_avatar 22d ago
Thanks! It was really cool to work a layer deeper in the stack
7.
by mips_avatar 22d ago
Direct link to the repo on github: https://github.com/jonready/vllm-ios
8.
vLLM-iOS: 88% Faster Multi-Agent Inference on iOS(jonready.com)
5 points by mips_avatar 22d ago|3 comments
9.
by mips_avatar 23d ago
I wouldn’t include YouTube given how only Google is allowed to index it
10.
by mips_avatar 23d ago
I hope it will force institutions to simplify entitlements. Right now the institutions can pretend like their kafkaesque system is fine because nobody manages to navigate the maze.
11.
by mips_avatar 23d ago
A single 3090 will train qwen 0.8B just fine. While it’s not a very capable model any training technique you would want to master can be used to make real progress. And all the skills you need to learn how to do this can be learned watching
12.
by mips_avatar 1mo ago
I think the biggest problem with the models is they don’t actually have any decent lookups except chunked document embedding search
13.
by mips_avatar 1mo ago
Depends on your pcie connection. If they're both x16 then it's pretty low overhead, x8 is ok, but x4 is too slow. Also it's a bit tricky getting an optimal setups with mismatched vram, I think you could probably still make
14.
by mips_avatar 1mo ago
Unfortunately AMD bought them, so I don't think we will get to see another release from them.
15.
by mips_avatar 1mo ago
Qwen3.5 was awesome: fairly open and fully featured. 3.8 lacking vision, nerfing thinking modes, and low context length feels pointless.
16.
by mips_avatar 1mo ago
It was inspiring watching the Gemini 3 pro team launch the model and bike away into the sunset never to launch anything again
17.
by mips_avatar 2mo ago
Like I have a list of a few hundred osm places websites that are clearly scam sites. I should be going one by one and filing them manually but I found this via a spam filter and it’s very robust. I should have a way of getting these scam si
18.
by mips_avatar 2mo ago
I’m grateful that a principled group of people run OSM. Much like I’m grateful that a principled group run Wikipedia. But the rigidness has costs that I don’t think are being appreciated.
19.
by mips_avatar 2mo ago
I appreciate OSM for maintaining a higher data quality bar than other projects (Overture places are mostly junk outside of USA), but it's also just artificially limiting itself by not allowing streamlined paths to data contributions.
20.
by mips_avatar 2mo ago
Would be interesting to see how fast it would be on 4x mac studio 512gb machines.
21.
by mips_avatar 2mo ago
Ok but a task that works fine on qwen 397b can be finetuned on qwen9b. But in every case so far when building the eval for evaluating the traces I’ve discovered a better prompt that closes the gap better than the finetuning.
22.
by mips_avatar 2mo ago
The problem i've had with finetuning models is that most of the time better prompting beats finetuning
23.
by mips_avatar 2mo ago
Problem is right now the biggest GPU boxes they have is single rtx pro 6000s.
24.
by mips_avatar 2mo ago
I’m still kind of shocked that Dean Ball can tweet such incendiary stuff about OpenAI policy. Like presumably OpenAI would prefer it if their staff don’t pick fights with Trump administration officials.
25.
by mips_avatar 2mo ago
I'm not sure it would be more fun, but you could probably ingest openstreetmaps data so it's actually a real manhattan
26.
by mips_avatar 2mo ago
So like oftentimes the picture will be of a church and there’s geographic coordinates for where the photo was taken. My qwen will use the geocoder to search for “church” at the coordinates of the photo and then read the Wikipedia articles a
27.
by mips_avatar 2mo ago
It's definitely being pushed more by the big labs recently, and it's interesting how they favor local hardware.
28.
by mips_avatar 2mo ago
Maybe I'm slow but I didn't use them much until recently when Cursor and Claude Code made them a main part of the harness.
29.
Agent swarms are great for local AI(jonready.com)
5 points by mips_avatar 2mo ago|4 comments
30.
by mips_avatar 2mo ago
Very few people travel to visit the Amazon especially in Peru, Bolivia, and Ecuador. In the absence of tourist money these amazing ecosystems are being turned into agriculture and logged. In the protected areas supported by eco tourism the

More ›