semessier

7 ms·

31.
by semessier 1y ago
Nvidia products including from the GPU/CUDA libraries world, the NICs and switches tend to feel like MVP frequently. It works in some cases, hopefully in the end but they are far from polished products without rough edges.
32.
by semessier 1y ago
adding to the question, ruling out fine-tuning for practicality, what about injecting names towards the embedding but not into the context?
33.
by semessier 1y ago
this could become a major thing in coarray world, but the issues start already: > ...Note that this does not support tensor engine, which is tied to CUDA and RDMA (via ibverbs). I.e. yet another CUDA married approach: the issue is not ib
34.
by semessier 1y ago
should post again when having 5% of the features of the other parallel file systems starting with RDMA, whereby it's not clear if this FS does even stripe that is if it is even a parallel file system
35.
by semessier 1y ago
which begs the question that which daily mouthwash is a) broadband effective against many including the right pathogens c) long-time, safe/tolerable for long time daily use. Candidate being Chlorhexidine 0.06% - are there opinions on i
36.
by semessier 1y ago
wondering how newborns under respiratory distress can survive hypoxia almost 30' post-birth with 50 and 57% measured SpO2, respectively.
37.
by semessier 1y ago
IBM hasn't had a hardware winner in more than 30 years, so this is still a pretty good record.
38.
by semessier 1y ago
still looking for vLLM to support Mac ARM Metal GPUs
39.
by semessier 1y ago
separate contexts for dev/test case development/documentation?
40.
by semessier 1y ago
ROCm, PyTorch?
41.
by semessier 1y ago
I had bet that matmult would be in transformer-optimized hardware costing a fraction of GPUs first class in torch 2 years ago with no reason to use GPUs any more. Wrong.
42.
by semessier 1y ago
what is the effective tax rate for Apple in Europe last year?
43.
by semessier 1y ago
legal tech apps via AI
44.
by semessier 1y ago
well, related: fractional GPUs to multiplex workloads for aggregate utilization have been a topic for some time with no definite (NVIDIA) solutions for it: https://vhpc.org
45.
by semessier 2y ago
this was/is the chip opportunity of the century. Even more optimized than the still general purpose nvidia cards. And no matrix mult is abstracted away for decades, don't need CUDA. So a chip would likely be much much easier than
46.
by semessier 2y ago
I don't get why more than 2 years after the ChatGPT release moment there is not a plethora of high-mem matrix-matrix matrix-vector hardware for the high-end available. Both high bandwidth and commodity DRAM. Both operations are very we
47.
by semessier 2y ago
the posting account was created 20 minutes ago
48.
by semessier 2y ago
The EU has actually been one of history's most successful projects for ensuring peace and prosperity in Europe. Looking at objective metrics like GDP growth, living standards, and decades of peace between former neighboring country adv
49.
by semessier 2y ago
no FEEP ion thrusters on Starlink
50.
by semessier 2y ago
amazing pieces of machinery the IBM RS/6000
51.
by semessier 2y ago
that resonates - less facts and more reasoning training data. The most low hanging in terms of non synthetic data probably being mathematical proofs. With prolog and the like many alternate reasoning paths could be generated. It's hard
52.
IBM helped us find our big-company footing(buildingslack.com)
2 points by semessier 2y ago|1 comments
53.
by semessier 2y ago
IBM used to solve customer problems with desired products and system integration. What happened - there is no hardware product success since the RS/6000 (1991) despite enormous annual R&D spending. Now here IBM is the customer and
54.
by semessier 2y ago
I wanted to ask if they do telnet/finger also, but there is no email listed.
55.
Ask HN: When did OpenAI first observe emergent reasoning and generalization?
2 points by semessier 2y ago|2 comments
56.
by semessier 3y ago
I believe the CentOS EoL will impact the re-builders albeit this complex. Not everything is GPL, they need the sources that may be a huge issue. To a lesser extent, probably the build script in the source rpem helps.
57.
by semessier 3y ago
I was looking for the decade first in the post - it fits the early 90s
58.
by semessier 3y ago
Apple has long been heralded for its unparalleled customer service, with the US Genius Bar often upheld as the gold standard in retail support. Yet beneath the sleek veneer of "genius" assistance lies a script — a carefully orches
59.
The Apple Store Blueprint: Genius Training or Psychological Manipulation Script?(gizmodo.com)
3 points by semessier 3y ago|2 comments
60.
by semessier 3y ago
is there something like open source / github type law opinion collaboration: case-in-point being Apple not paying its taxes being about to get out of more than $14 billion with a last appeal needing arguments.

More ›