this post was submitted on 09 May 2025
179 points (96.9% liked)

Technology

69891 readers
2558 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots

founded 2 years ago
MODERATORS
L3s@lemmy.world
enu@lemmy.world
technopagan@lemmy.world
L4s@lemmy.world
L3s@hackingne.ws
L4s@hackingne.ws
179
Microsoft Bans Employees From Using DeepSeek App (www.sunrisegeek.com)
submitted 1 day ago by helloworld7@lemmy.world to c/technology@lemmy.world
12 comments fedilink hide all child comments
top 12 comments
sorted by: hot top controversial new old
[–] brucethemoose@lemmy.world 30 points 1 day ago* (last edited 1 day ago) (1 children)

One can get Deepseek R1 from many providers (including US hosts, or various other nationalities). Microsoft even has their own anti-CCP finetune, MIT licensed: https://huggingface.co/microsoft/MAI-DS-R1

...Banning the app is reasonable, and a tiny inconvenience for anyone who needs DS.

In other words, this is a big nothingburger because V3/R1 are open models. The story would be different if it was (say) an API-only model like Qwen Max or GPT4o, where ultimately one is beholden to the trainer's servers.

[–] ilmagico@lemmy.world 3 points 1 day ago (1 children)

I literally run deepseek r1 on my laptop via ollama, and many other models, nothing gets sent to anybody. Granted, it's the smaller 7b parameter model, but still plenty good.

Microsoft could easily host the full model on their infrastructure if they needed it.

[–] brucethemoose@lemmy.world 5 points 1 day ago (1 children)

True, though there's a big output difference between the 7B distil (or even 32B/70B) and the full model.

And Microsoft does host R1 already, heh. Again, this headline is a big nothingburger.

Also (random aside here), you should consider switching from ollama. They're making some FOSS unfriendly moves, and depending on your hardware, better backends could host 14B models at longer context, and similar or better speeds.

[–] Sabata11792@ani.social 3 points 1 day ago (1 children)

What other back ends are good?

[–] brucethemoose@lemmy.world 4 points 1 day ago* (last edited 1 day ago)

Completely depends on your laptop hardware, but generally:

  • TabbyAPI (exllamav2/exllamav3)
  • ik_llama.cpp, and its openai server
  • kobold.cpp (or kobold.cpp rocm, or croco.cpp, depends)
  • An MLX host with one of the new distillation quantizations
  • Text-gen-web-ui (slow, but supports a lot of samplers and some exotic quantizations)
  • SGLang (extremely fast for parallel calls if thats what you want).
  • Aphrodite Engine (lots of samplers, and fast at the expense of some VRAM usage).

I use text-gen-web-ui at the moment only because TabbyAPI is a little broken with exllamav3 (which is utterly awesome for Qwen3), otherwise I'd almost always stick to TabbyAPI.

Tell me (vaguely) what your system has, and I can be more specific.

[–] MeaanBeaan@lemmy.world 20 points 1 day ago (2 children)

Correct me if I'm wrong here but a company can't dictate what you do on your own time on your own hardware, so I assume this simply affects work computers. Assuming thats the case I don't really see a problem here. I've never been able to download any applications at all on any work computer I've ever used short of apps the company itself uses.

Seems completely understandable to me to bar employees from using a competing service especially if there are genuine security concerns.

[–] paraphrand@lemmy.world 12 points 1 day ago (1 children)

Ever heard about software devs not being able to work on personal projects because all the code they produce, even off the clock is owned by their employer?

[–] ilmagico@lemmy.world 1 points 1 day ago

Hasn't been true for my past two jobs at least (US based), what I do outside of company premises / my own hardware and my own time is mine. They only own what was done on company's dime. Not saying it doesn't happen, but that's not my experience so far, and I'm not sure if would be legal.

[–] shiroininja@lemmy.world 5 points 1 day ago

I mean if It’s only for business machines I get it. If you want a proper silo where your ip and isn’t going to be stolen by an LLM, any organization should run an in house self hosted model that is trained on their own data and doesn’t pass data back to the upstream. That just makes sense. Especially if it just passes your work to your competitors.

I mean if we ever properly gleam information from LLMs, it’ll be the biggest source of leaks and whistleblowing ever created from video games to National security. Don’t use cloud hosted llms if you want privacy and security.

[–] Tarquinn2049@lemmy.world 5 points 1 day ago (1 children)

Ignoring the content of the article for a second. Did they not proofread this at all? There are so many spelling, grammar, and sentence fragment mistakes. I would joke that it was written by AI, but it has too many mistakes for even that to be the case.

[–] kevin2107@lemmy.world 5 points 1 day ago

Atleast we know its by humans

[–] Vaie@lemm.ee -1 points 1 day ago

AI has unreliable results for all but morons who can’t tell the difference due to lacking basic reading comprehension skills.

The fact that its essays and other output are considered at all “good” is an indicator of how poor our education system really is. Most are objectively bad.

That said, I’m far more comfortable with China getting my data from it than the USA. But there’s zero reason for me to use it in the first place.