uɐɥdǝʇS
@StephanSturges
Followers
5K
Following
45K
Media
6K
Statuses
24K
UAVs are cool i guess
null island
Joined October 2008
I'm finally getting around to releasing the Unidrone FOSS AI models for drones, check them out here: https://t.co/LlJkTl0GRi Sorry for the delayed FOSS release, it's hard to find any free time these days to work on these kinds of things... anyway let me know what you think 😄
11
61
338
RF-DETR paper is finally on arXiv - real time detection with DINOv2 backbone - runs neural architecture search (NAS) over about 6000 architecture variants - uses weight sharing across all configs - first real-time segmentation DETR to break past top YOLO results ↓ more
13
55
476
GPT-5 is in 35th position out of 67 AI models based on "user satisfaction" in this French-government-created report. https://t.co/j5BS2PliZA Not sure what this means, but there you go 😄
comparia.beta.gouv.fr
compar:IA est un outil permettant de comparer à l’aveugle différents modèles d'IA conversationnelle pour sensibiliser aux enjeux de l'IA générative (biais, impact environmental) et constituer des...
1
0
1
Hey @OpenAI this happened again
It's incredibly annoying how @ChatGPTapp keeps reverting to "auto" even when I keep re-selecting "pro mode" in the same discussion. I don't want "auto" mode. I never want "auto" mode. Please stop giving me "auto" mode.
3
0
1
It's incredibly annoying how @ChatGPTapp keeps reverting to "auto" even when I keep re-selecting "pro mode" in the same discussion. I don't want "auto" mode. I never want "auto" mode. Please stop giving me "auto" mode.
1
0
8
How it started (2018): - Bought a laser - Developed instant quoting system - Hired 3 people - We can cut metal with lasers! - We can ship it to your door! How it went for 7 years (from what I can remember): - Bought another laser - Hired 9 more people - Added 23k lines of code
171
215
5K
repeat after me: commoditize your complement expect a barrage of consumer electronics leveraging LLMs / VLMs coming out of China over the next 10 years
The gap is closing. China is catching up. Kimi-K2 Thinking crushes GPT-5 and Claude 4.5 Sonnet in several benchmarks, while costing 6 times less compared to Sonnet It's the best open-source model period Its core focus is on agentic tasks and software development. It can now
0
1
2