Hi,
On 2025/03/14 14:22, Sam James wrote:
Arsen Arsenović writes:
Filip Kobierski writes:
On Monday, March 10th, 2025 at 21:40, Alfredo Tupone wrote:
To declutter sci-libs and dev-libs from most of the "so called"
AI packages I think that a new category should be created.
Maybe sci-ai/ o
> On Sat, 15 Mar 2025, Andreas K Huettel wrote:
> ++ for sci-ai
> (Now, is it really sci-ai, or should we also come up with dev-ai (for
> libraries without explicit scientific context) and sys-ai (for accelerator
> device drivers) in addition? :)
The possibility that we could later add a de
Gordon Pettey writes:
> IMHO, "ai" is an extremely overloaded and over- and mis-used term.
> It's nothing but glorified pattern matching, and calling everything "ai"
> is very buzzwordy. I'd much rather see it named "ml".
Personally, I don't really care that a relatively well-understood word
has
Arsen Arsenović writes:
> Filip Kobierski writes:
>
>> On Monday, March 10th, 2025 at 21:40, Alfredo Tupone
>> wrote:
>>> To declutter sci-libs and dev-libs from most of the "so called"
>>> AI packages I think that a new category should be created.
>>> Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep
>
> This appears to leave us with sci-ai/* because:
>
> First, 'AI' seems to be the term that is commonly used (just look at
> this mail's subject) and understood.
>
> Secondly, while others may find sci-ai to buzzwordy, that could also
> been seen as an advantage.
This.
Buzzwordy is kinda
On Wed, Mar 12, 2025 at 08:47:42AM +0100, Florian Schmaus wrote:
> On 10/03/2025 21.40, Alfredo Tupone wrote:
> > To declutter sci-libs and dev-libs from most of the "so called"
> > AI packages I think that a new category should be created.
> > Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
>
IMHO, "ai" is an extremely overloaded and over- and mis-used term.
It's nothing but glorified pattern matching, and calling everything "ai"
is very buzzwordy. I'd much rather see it named "ml".
On Fri, Mar 14, 2025 at 7:23 AM Sam James wrote:
> Arsen Arsenović writes:
>
> > Filip Kobierski wri
Filip Kobierski writes:
> On Monday, March 10th, 2025 at 21:40, Alfredo Tupone
> wrote:
>> To declutter sci-libs and dev-libs from most of the "so called"
>> AI packages I think that a new category should be created.
>> Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
>
> I really like thi i
On 3/10/25 6:59 PM, Maciej Barć wrote:
> W dniu 10.03.2025 o 23:42, Eli Schwartz pisze:
>> I don't understand your argument at all. "ml" is hardly a reserved
>> concept, and dev-ml exists precisely for "libraries and utilities
>> relevant to the ML programming language", which isn't going to get
>>
On Wed, 12 Mar 2025 08:47:42 +0100
Florian Schmaus wrote:
> This appears to leave us with sci-ai/* because:
>
> First, 'AI' seems to be the term that is commonly used (just look at
> this mail's subject) and understood.
>
> Secondly, while others may find sci-ai to buzzwordy, that could also
On Mon, 10 Mar 2025 18:42:49 -0400
Eli Schwartz wrote:
> On 3/10/25 4:53 PM, Maciej Barć wrote:
> > Hi!
> >
> >> Although maybe it should be sci-ml.
> >
> > Let's _not_ use *-ml since for us ml stands for OCaml (which comes from
> > ML - "Meta langauge").
> >
> > sci-ai, dev-ai, and app-ai
On 10/03/2025 21.40, Alfredo Tupone wrote:
To declutter sci-libs and dev-libs from most of the "so called"
AI packages I think that a new category should be created.
Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
Thanks for your proposal.
I would go with sci-ai/*, even if all packages und
W dniu 10.03.2025 o 23:42, Eli Schwartz pisze:
I don't understand your argument at all. "ml" is hardly a reserved
concept, and dev-ml exists precisely for "libraries and utilities
relevant to the ML programming language", which isn't going to get
confused with sci-ml/ for the same reason nobody w
To declutter sci-libs and dev-libs from most of the "so called"
AI packages I think that a new category should be created.
Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
sci-libs has now 264 packages
The packages that I can move from sci-libs in the new category are:
caffe2
datasets
evaluate
On 3/10/25 4:53 PM, Maciej Barć wrote:
> Hi!
>
>> Although maybe it should be sci-ml.
>
> Let's _not_ use *-ml since for us ml stands for OCaml (which comes from
> ML - "Meta langauge").
>
> sci-ai, dev-ai, and app-ai (say, "app-ai/ollama"?) are nice IMO.
- please don't top-post
- Let's _not_
> On Mon, 10 Mar 2025, Maciej Barć wrote:
> W dniu 10.03.2025 o 23:42, Eli Schwartz pisze:
>> I don't understand your argument at all. "ml" is hardly a reserved
>> concept, and dev-ml exists precisely for "libraries and utilities
>> relevant to the ML programming language", which isn't going t
On Mon, Mar 10, 2025 at 09:40:09PM +0100, Alfredo Tupone wrote:
> To declutter sci-libs and dev-libs from most of the "so called"
> AI packages I think that a new category should be created.
> Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
> The packages that I can move from dev-libs in the
On Monday, March 10th, 2025 at 21:40, Alfredo Tupone wrote:
> To declutter sci-libs and dev-libs from most of the "so called"
> AI packages I think that a new category should be created.
> Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
I really like thi idea.
For better or worse the field is
Hi!
Although maybe it should be sci-ml.
Let's _not_ use *-ml since for us ml stands for OCaml (which comes from
ML - "Meta langauge").
sci-ai, dev-ai, and app-ai (say, "app-ai/ollama"?) are nice IMO.
W dniu 10.03.2025 o 21:49, Eli Schwartz pisze:
On 3/10/25 4:40 PM, Alfredo Tupone wrote:
On 3/10/25 4:40 PM, Alfredo Tupone wrote:
> To declutter sci-libs and dev-libs from most of the "so called"
> AI packages I think that a new category should be created.
> Maybe sci-ai/ or dev-ai/ or sci-dl/ (deep-learning)
Of the three I favor sci-dl, since by and large these aren't really
about
20 matches
Mail list logo
