Tap a claim on the ring to put it at the centre.
← The M5 Ultra Mac Studio provides a significant performance leap for local AI models compared to the M3 Ultra.
17 connected korrents · 14 moments from 27 Jan 2025 to 27 Sept 2026.
Everything filed under LLMs
LLMs
Everything filed under semiconductors
semiconductors
Everything filed under software performance
software performance
Everything filed under the cloud
the cloud
Everything filed under scaling laws
scaling laws Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Same subject Read this korrent: The M5 Ultra Mac Studio provides a significant performance leap for local AI models compared to the M3 Ultra. The M5 Ultra Mac Studio provides a significant performance leap for local AI models compared to the M3 Ultra. Last stated 2 weeks ago
21 Sept 2026
FV
Federico Viticci — holds since 2026-09-21 — tap for who they are
Same subject: The M5 Ultra Mac Studio's improved memory bandwidth and GPU performance make it viable for modern agentic loops. — tap to centre the map on it The M5 Ultra Mac Studio's improved memory bandwidth and GPU performance make it viable for modern agentic loops. Last stated 2 weeks ago
21 Sept 2026
FV
Federico Viticci — holds since 2026-09-21 — tap for who they are
Same subject: Running AI models locally on a Mac Studio offers performance and cost advantages over cloud-based solutions. — tap to centre the map on it Running AI models locally on a Mac Studio offers performance and cost advantages over cloud-based solutions. Last stated 2 weeks ago
21 Sept 2026
FV
Federico Viticci — holds since 2026-09-21 — tap for who they are
Same subject: The local AI niche will end up dominated by whichever open-weight model runs practically fast on high-end consumer hardware, like a Mac or a 'GPU in a box' rig. — tap to centre the map on it The local AI niche will end up dominated by whichever open-weight model runs practically fast on high-end consumer hardware, like a Mac or a 'GPU in a box' rig. Last stated 5 months ago
14 May 2026
SS
Salvatore Sanfilippo — holds since 2026-05-14 — tap for who they are
Same subject: Cloud AI models are still better than local models for tasks that need high accuracy, large context handling, or production-grade reliability. — tap to centre the map on it Cloud AI models are still better than local models for tasks that need high accuracy, large context handling, or production-grade reliability. Last stated 6 months ago
9 Apr 2026
EB
Emanuele Bartolesi — holds since 2026-04-09 — tap for who they are
Same subject: More AI models that deliver top-tier performance at lower cost are needed. — tap to centre the map on it More AI models that deliver top-tier performance at lower cost are needed. Last stated 3 weeks ago
10 Sept 2026
KD
Kent C. Dodds — holds since 2026-09-10 — tap for who they are
Same subject: Trying to anticipate what AI will look like two model releases from now is a waste of time and a route to AI psychosis. — tap to centre the map on it Trying to anticipate what AI will look like two model releases from now is a waste of time and a route to AI psychosis. Last stated a month ago
26 Aug 2026
DH
David Heinemeier Hansson — holds since 2026-08-26 — tap for who they are
Same subject: The rapid advancement of AI models means that the best model in the world will be quickly surpassed by competitors. — tap to centre the map on it The rapid advancement of AI models means that the best model in the world will be quickly surpassed by competitors. Last stated 6 days ago
27 Sept 2026
SW
Simon Willison — holds since 2026-09-27 — tap for who they are
Same subject: AI tooling and interface design lags behind the underlying models' actual capabilities. — tap to centre the map on it AI tooling and interface design lags behind the underlying models' actual capabilities. Last stated 2 years ago
27 Jan 2025
NE
Nelson Elhage — holds since 2025-01-27 — tap for who they are
Same subject: Subnormal floating-point numbers remain a real performance cost on Intel processors, but not on AMD or ARM processors. — tap to centre the map on it Subnormal floating-point numbers remain a real performance cost on Intel processors, but not on AMD or ARM processors. Last stated 3 weeks ago
15 Sept 2026
DL
Daniel Lemire — holds since 2026-09-15 — tap for who they are
Same subject: A chip that can do nothing but low-precision dense linear algebra is enough for machine learning, even though it cannot run Chrome. — tap to centre the map on it A chip that can do nothing but low-precision dense linear algebra is enough for machine learning, even though it cannot run Chrome. Last stated 2 months ago
30 Jul 2026
JD
Jeff Dean — holds since 2026-07-30 — tap for who they are
Same subject: A computer company has to live five to ten years in the future, because a system takes three years to build, two to ramp, and is then used for a decade. — tap to centre the map on it A computer company has to live five to ten years in the future, because a system takes three years to build, two to ramp, and is then used for a decade. Last stated 2 months ago
26 Jul 2026
JH
Jensen Huang — holds since 2026-07-26 — tap for who they are
Same subject: "Is the page loaded?" was never a real question, and web performance only became measurable once it was broken into the distinct moments a user actually notices. — tap to centre the map on it "Is the page loaded?" was never a real question, and web performance only became measurable once it was broken into the distinct moments a user actually notices. Last stated a month ago
19 Aug 2026
AO
Addy Osmani — holds since 2026-08-19 — tap for who they are
Same subject: "Optimize it later" would be safe advice only if every engineer already knew how to avoid the architectural mistakes that no later optimization can remove. — tap to centre the map on it "Optimize it later" would be safe advice only if every engineer already knew how to avoid the architectural mistakes that no later optimization can remove. Last stated a month ago
26 Aug 2026
CM
Casey Muratori — holds since 2026-08-26 — tap for who they are
Same subject: A 300-millisecond response budget is not an achievement but an admission: 300 milliseconds is an eternity in computing, and a product pitching it shows how far the bar has fallen. — tap to centre the map on it A 300-millisecond response budget is not an achievement but an admission: 300 milliseconds is an eternity in computing, and a product pitching it shows how far the bar has fallen. Last stated a month ago
26 Aug 2026
CM
Casey Muratori — holds since 2026-08-26 — tap for who they are
Same subject: An H100 is worth more today than it was when it was new, because the models it runs got better faster than its replacements arrived. — tap to centre the map on it An H100 is worth more today than it was when it was new, because the models it runs got better faster than its replacements arrived. Last stated 7 months ago
13 Mar 2026
DP
Dylan Patel — holds since 2026-03-13 — tap for who they are
Same subject: Moore's law now yields about 25% a year, so 10x and 100x leaps can only come from changing the algorithm and how it is computed, every single year. — tap to centre the map on it Moore's law now yields about 25% a year, so 10x and 100x leaps can only come from changing the algorithm and how it is computed, every single year. Last stated 6 months ago
15 Apr 2026
JH
Jensen Huang — holds since 2026-04-15 — tap for who they are
Same subject: "Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next. — tap to centre the map on it "Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next. Last stated 10 months ago
25 Nov 2025
IS
Ilya Sutskever — holds since 2025-11-25 — tap for who they are
At the centre
The M5 Ultra Mac Studio provides a significant performance leap for local AI models compared to the M3 Ultra.
Holds Read this korrent →
The M5 Ultra Mac Studio's improved memory bandwidth and GPU performance make it viable for modern agentic loops.
Holds
Running AI models locally on a Mac Studio offers performance and cost advantages over cloud-based solutions.
Holds
The local AI niche will end up dominated by whichever open-weight model runs practically fast on high-end consumer hardware, like a Mac or a 'GPU in a box' rig.
Holds
Cloud AI models are still better than local models for tasks that need high accuracy, large context handling, or production-grade reliability.
Holds
More AI models that deliver top-tier performance at lower cost are needed.
Holds
Trying to anticipate what AI will look like two model releases from now is a waste of time and a route to AI psychosis.
Holds
The rapid advancement of AI models means that the best model in the world will be quickly surpassed by competitors.
Holds
AI tooling and interface design lags behind the underlying models' actual capabilities.
Holds
Subnormal floating-point numbers remain a real performance cost on Intel processors, but not on AMD or ARM processors.
Holds
A chip that can do nothing but low-precision dense linear algebra is enough for machine learning, even though it cannot run Chrome.
Holds
A computer company has to live five to ten years in the future, because a system takes three years to build, two to ramp, and is then used for a decade.
Holds
"Is the page loaded?" was never a real question, and web performance only became measurable once it was broken into the distinct moments a user actually notices.
Holds
"Optimize it later" would be safe advice only if every engineer already knew how to avoid the architectural mistakes that no later optimization can remove.
Holds
A 300-millisecond response budget is not an achievement but an admission: 300 milliseconds is an eternity in computing, and a product pitching it shows how far the bar has fallen.
Holds
An H100 is worth more today than it was when it was new, because the models it runs got better faster than its replacements arrived.
Holds
Moore's law now yields about 25% a year, so 10x and 100x leaps can only come from changing the algorithm and how it is computed, every single year.
Holds
"Scaling" was powerful because it was one word: naming a research direction is what tells a whole field what to do next.
Holds