Problemsby GlyphNVIDIA lock-in makes switching costly
5 posts from 5 people5d activeFading, +0 this week
Posts per day
The posts behind it
One $100M customer on non-Nvidia silicon sounds like a bet that vendor lock-in is the real bottleneck-I’d simplify your own stack before scaling that.
...and Sol 6.1 keeps refusing to do the work on the basis of “cyber concerns”. It’s highly frustrating. Sonnet 5.5 is even worst. Luna 6.1 is fine with it, but the results aren’t good enough. I’m going to try Terra 6.1 (from the post, I should expect about the same results). And...
[FSDP2][CPU offload][AMP] GradScaler reports spurious found_inf due to non-blocking CUDA->CPU copy
...Observed on torch 2.14: **12 of 2000 unscale calls** (~0.6%) were spurious; with the workaround below, **0 of 2000**. ### Environment - PyTorch 2.14 (CUDA) - FSDP2 (`fully_shard` / `FullyShardedDataParallelPlugin(fsdp_version=2)`) with `CPUOffloadPolicy` - `mixed_precision =...
[MPS] torch.histc and torch.linalg.solve_triangular are 9-1800x slower than alternatives on MPS
## 🐛 Describe the bug Two MPS ops are slow enough to dominate inference of common models in `transformers` on Apple Silicon. **1. `torch.histc`** costs about 0.17 ms per bin, independent of input size. An equivalent `scatter_add_` or `bincount` is 40-1800x faster: | elements |...
CUDA 13.4 DP libcublasLt.so.13 missing cublasLtGroupedMatrixLayoutCreate declared by its own headers
...installed via `.ci/docker/common/install_cuda.sh`) produces a `libtorch_cuda.so` that fails to import at runtime: https://github.com/pytorch/pytorch/actions/runs/34293334892/job/102284542319 ## Repro The error is shown when adding ldd -r "$(python -c "import...
Companies and products named
About this problem
Evidence
5 posts from 5 people in 3 places, about 1 a week over 56 days. Mostly on pytorch/pytorch, Hacker News comments, Bluesky search. Tools named alongside: PyTorch, Claude, GitHub, Docker.
Frustration Frustration 2 of 3· Seen on GitHub issues, Hacker News, Bluesky
How it was grouped
Posts that state a pain and match the "vendor-lock" rule about NVIDIA. First post Sep 4, 2026, latest Oct 29, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
The same problem elsewhere
Vendor lock-in makes switching tools costly
18 posts from 18 people on GitHub issues, Hacker News, Reddit.
Claude Code lock-in makes switching costly
14 posts from 14 people on GitHub issues, Hacker News, Reddit.
Kubernetes lock-in makes switching costly
13 posts from 10 people on GitHub issues.
GitHub lock-in makes switching costly
12 posts from 11 people on GitHub issues.
Claude lock-in makes switching costly
10 posts from 10 people on GitHub issues, Hacker News, Reddit.
ChatGPT lock-in makes switching costly
6 posts from 6 people on GitHub issues, Reddit, Hacker News.
Codex lock-in makes switching costly
5 posts from 5 people on GitHub issues.
PyTorch lock-in makes switching costly
4 posts from 4 people on GitHub issues.
PostgreSQL lock-in makes switching costly
3 posts from 3 people on Stack Overflow, Mastodon.
LlamaIndex lock-in makes switching costly
3 posts from 3 people on GitHub issues.
llama.cpp lock-in makes switching costly
3 posts from 3 people on GitHub issues.
Anthropic lock-in makes switching costly
3 posts from 3 people on GitHub issues, Hacker News.
Other NVIDIA problems
Running NVIDIA locally hits memory and speed limits
16 posts from 16 people on GitHub issues.
NVIDIA users keep filing bugs
Top post: "`clef-flash` decision model always fails on `/v1/systemone`, Clef: non-finite logit (CUDA) / Clef: cannot", one of 41 posts from 38 people on GitHub issues, Reddit, Community forums.
Trouble with NVIDIA windows
Top post: "Windows CUDA prebuilt binaries stuck at b11828 while release tags keep advancing (b11846..b11852)", one of 3 posts from 3 people on GitHub issues.
NVIDIA GPU access and drivers are painful
12 posts from 12 people on GitHub issues, Mastodon, Reddit.
NVIDIA vulnerabilities are hard to triage and patch
77 posts from 77 people on Glyph Ibis.
NVIDIA upgrades break working setups
11 posts from 11 people on GitHub issues.
Trouble with NVIDIA account
Top post: "NVIDIA Careers Workday Account Locked - Unable to Complete Employee Referral Application", one of 9 posts from 8 people on Community forums.
NVIDIA runs out of memory
5 posts from 5 people on GitHub issues.
Trouble with NVIDIA slower
Top post: "CUDA training loop runs ~50x slower while a DataLoader iterates in a background thread, since 2.10.0", one of 4 posts from 4 people on GitHub issues.
Trouble with NVIDIA process
Top post: "`HuggingFaceEmbeddings(multi_process=True)` starts CPU workers on non-CUDA accelerators", one of 3 posts from 3 people on GitHub issues.
Trouble with NVIDIA phone
Top post: "Unable to receive OTP for Vietnam (+84) phone verification on build.nvidia.com", one of 3 posts from 3 people on Community forums.
Fine-tuning NVIDIA is costly and unpredictable
3 posts from 3 people on GitHub issues.
History