banner
oliverguhr.bsky.social
@oliverguhr.bsky.social
I build voice interfaces for robots. #ai #voice #robots
Porsche kehrt zum Verbrenner zurück, während chinesische Hersteller elektrisch von Rekord zu Rekord jagen.
September 22, 2025 at 7:16 AM
The only thing that is certain, when travelling with @deutschebahn.com - you will be to late 😆
July 1, 2025 at 3:48 PM
Reposted
VEO 3

Prompt: Generate a TV news anchor with a British accent, the anchor says "in shocking news, Jake Kaye Rowe ling's yacht sank with her on board after being attacked by orcas off the coast of turkey" include the announcement in the dialogue.

This took one prompt and 2 minutes to generate.
May 21, 2025 at 2:26 AM
Is the hype real?
May 5, 2025 at 7:23 AM
Reposted
Spectacular view of 100,000 people attending the #HandsOff march in NYC
April 5, 2025 at 8:44 PM
Reposted
#jobalert 🚨 We have 34 (!) open PhD positions @scadsai.bsky.social in Dresden and Leipzig (Germany).
The cross-disciplinary topics range from NLP, KG, Protein design, Math foundations, Visualization, Ethics, ML, RL, DL, all things #AI. Join us! 💻🤖🚀
www.verw.tu-dresden.de/StellAus/dow...
March 13, 2025 at 8:27 AM
„Nvidia needs to grow at least 30% a year on average for the next decade to justify its current share price. Any sign of disruption to that trajectory could lead to a sharp selloff.“ 🤔
February 27, 2025 at 7:21 AM
Ups, looks like huggingface is offline.
February 3, 2025 at 1:28 PM
I got 1000€ in Azure credits to spend in the next two days. Any cool ideas what we can do with it?
January 29, 2025 at 12:48 PM
Docling is an easy to use python lib that converts PDFs, PowerPoints and more to markdown. I tested it today and the results looked quite good.

github.com/DS4SD/docling
GitHub - DS4SD/docling: Get your documents ready for gen AI
Get your documents ready for gen AI. Contribute to DS4SD/docling development by creating an account on GitHub.
github.com
November 19, 2024 at 7:48 PM
This documentation of an assault on Russian tranches by international legion in Ukraine is just crazy. It’s very well made and show the horrors of this war youtu.be/402hOvSZ7tA
The Chosen Company. Violence of Action
YouTube video by International Legion for the Defence of Ukraine
youtu.be
November 10, 2024 at 10:47 PM
Reposted
I know people like to shit on these LLM models, but these tools to improve writing are a boon for immigrants.
It is the ultimate accessibility tool.
September 9, 2024 at 6:08 PM
Reposted
Historikerin begeistert, dass sie live miterleben kann, wie das alles angefangen hat damals
Historikerin begeistert, dass sie live miterleben kann, wie das alles angefangen hat damals
Erfurt (dpo) - Dass sie so eine Chance noch zu Lebzeiten bekommt, hätte sie nicht gedacht: Die Historikerin Dr. Carina Maguhn aus Erfurt mit dem Spezi
www.der-postillon.com
September 2, 2024 at 11:45 AM
Immerhin hat #crowdstrike sein Werbeversprechen gehalten 😬
July 19, 2024 at 12:27 PM
📣Einen Tag vor den Europa- und Kommunalwahlen gehen wir wieder groß auf die Straßen für Demokratie und gegen Rechtsextremismus !

#WirSindDieBrandmauer #wirsinddiebrandmauerdresden #wirsindmehr #niewiederistjetzt #DD0806
June 8, 2024 at 1:47 PM
Wow, LLMs as interactive 3D models right in your browser.
bbycroft.net/llm
December 4, 2023 at 10:01 AM
Since GPUs are expensive and quite rare, first teams are moving to CPU-based AI inference. I looked into this as well for my research, and I found that CPUs can be a solid choice for many applications.
www.kedglobal.com/tech,-media-...
November 2, 2023 at 2:43 PM
Reposted
The 2023 state of the climate report: it's fucked

academic.oup.com/bioscience/a...
October 25, 2023 at 8:13 AM
Spotify raises the prices by 20%. Is justified or greed?
October 20, 2023 at 7:05 AM
Nice, Nvidias speech recognition model for German is as good as Whisper (in my tests) and it is much faster. At least 4x faster than Whisper small.
October 13, 2023 at 2:47 PM
Here is what Dalle-3 can do. I prompted the model with "Photos of a superhero programmer saving the world." and it came up with these pictures. Really clever is, that they generate a more refined prompt from your initial prompt. So basically, they automated the prompt engineering part.
October 13, 2023 at 7:39 AM
So cool! @huggingface.bsky.social lets you generate images free, fast and in high quality!
Try it at: huggingface.co/spaces/googl...
October 11, 2023 at 9:03 AM