Tap a claim on the ring to put it at the centre.
← What matters about AI hardware is essentially its cost: cost per flop and overall system cost.
17 connected korrents · 16 moments from 27 Mar 2023 to 31 Oct 2026.
Everything filed under semiconductors
semiconductors
Everything filed under energy
energy
Everything filed under AI and human skill
AI and human skill
Everything filed under ideas vs execution
ideas vs execution
Everything filed under design
design
Everything filed under simplicity
simplicity
Everything filed under Apple
Apple
Everything filed under AI sentience
AI sentience
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: What matters about AI hardware is essentially its cost: cost per flop and overall system cost.
What matters about AI hardware is essentially its cost: cost per flop and overall system cost.
Last stated 4 years ago
27 Mar 2023
IS
Ilya Sutskever — holds since 2023-03-27 — tap for who they are
Same subject: The numbers an AI engineer should know by heart are memory bandwidths and the energy cost of a single operation. — tap to centre the map on it
The numbers an AI engineer should know by heart are memory bandwidths and the energy cost of a single operation.
Last stated 4 months ago
30 Jul 2026
JD
Jeff Dean — holds since 2026-07-30 — tap for who they are
Same subject: Details, understanding and thinking still matter in programming. — tap to centre the map on it
Details, understanding and thinking still matter in programming.
Last stated 3 months ago
25 Aug 2026
NE
NeetCode — holds since 2026-08-25 — tap for who they are
Same subject: People do not value or care about something built by AI, regardless of its quality, because no personal effort went into it. — tap to centre the map on it
People do not value or care about something built by AI, regardless of its quality, because no personal effort went into it.
Last stated 2 months ago
21 Sept 2026
SJ
Stefan Judis — holds since 2026-09-21 — tap for who they are
Same subject: If you think through the customer experience you want to create, you end up realizing hardware is a necessity. — tap to centre the map on it
If you think through the customer experience you want to create, you end up realizing hardware is a necessity.
Last stated 7 months ago
12 Apr 2026
ES
Evan Spiegel — holds since 2026-04-12 — tap for who they are
Same subject: Cheap AI-built features still cost customers complexity and cognitive load. — tap to centre the map on it
Cheap AI-built features still cost customers complexity and cognitive load.
Last stated 3 months ago
2 Sept 2026
DS
Dharmesh Shah — holds since 2026-09-02 — tap for who they are
Same subject: A product's success depends on hardware and software working together, not hardware alone. — tap to centre the map on it
A product's success depends on hardware and software working together, not hardware alone.
Last stated 2 months ago
10 Sept 2026
MS
M.G. Siegler — holds since 2026-09-10 — tap for who they are
Same subject: AI capabilities improve daily, but hardware takes months and years to ship and build at scale. — tap to centre the map on it
AI capabilities improve daily, but hardware takes months and years to ship and build at scale.
Last stated 3 weeks ago
31 Oct 2026
MN
Marily Nika — holds since 2026-10-31 — tap for who they are
Same subject: Faultless computer hardware could still support confusion, forgetting or a sense of beauty — tap to centre the map on it
Faultless computer hardware could still support confusion, forgetting or a sense of beauty
DH
Douglas Hofstadter — holds since c. 1979 — tap for who they are
Same subject: A ban on self-recursive improvement in AI models would create a big incentive for some other parts of the world to invest a lot more in compute. — tap to centre the map on it
A ban on self-recursive improvement in AI models would create a big incentive for some other parts of the world to invest a lot more in compute.
Last stated 2 months ago
22 Sept 2026
TC
Tyler Cowen — holds since 2026-09-22 — 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 4 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 4 months ago
26 Jul 2026
JH
Jensen Huang — holds since 2026-07-26 — tap for who they are
Same subject: AI data centers have shifted from about 1 CPU per 8 GPUs to roughly 1:4, and the ratio could shrink to 1:1. — tap to centre the map on it
AI data centers have shifted from about 1 CPU per 8 GPUs to roughly 1:4, and the ratio could shrink to 1:1.
Last stated 2 months ago
24 Sept 2026
GO
Gergely Orosz — holds since 2026-09-24 — tap for who they are
Same subject: AI workloads are best served by clouds that bundle power, data center infrastructure and GPUs and strip out unnecessary software layers. — tap to centre the map on it
AI workloads are best served by clouds that bundle power, data center infrastructure and GPUs and strip out unnecessary software layers.
Last stated 2 months ago
8 Sept 2026
CP
Chamath Palihapitiya — holds since 2026-09-08 — tap for who they are
Same subject: Building capacity at massive scale on a single hardware generation is the trap; pacing matters more than gigawatts. — tap to centre the map on it
Building capacity at massive scale on a single hardware generation is the trap; pacing matters more than gigawatts.
Last stated a year ago
12 Nov 2025
SN
Satya Nadella — holds since 2025-11-12 — 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 8 months ago
13 Mar 2026
DP
Dylan Patel — holds since 2026-03-13 — tap for who they are
Same subject: Around 2030 to 2032 AI progress has to taper as compute runs out, unless very powerful capabilities arrive or fabs and investment scale faster than expected. — tap to centre the map on it
Around 2030 to 2032 AI progress has to taper as compute runs out, unless very powerful capabilities arrive or fabs and investment scale faster than expected.
Last stated a year ago
8 Jul 2025
RG
Ryan Greenblatt — holds since 2025-07-08 — 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 7 months ago
15 Apr 2026
JH
Jensen Huang — holds since 2026-04-15 — tap for who they are
same subject or similar wording a cloud: claims about one subject, named for it bar: when it was last stated, on a scale from 2015 to today — full is today a face: someone who holds the claim — tap it for who they are
At the centre
What matters about AI hardware is essentially its cost: cost per flop and overall system cost.
Last stated 27 Mar 2023 · 4 years ago
Holds Ilya Sutskever
Read this korrent →
Similar wording
The numbers an AI engineer should know by heart are memory bandwidths and the energy cost of a single operation.
Last stated 30 Jul 2026 · 4 months ago
Holds Jeff Dean
Similar wording
Details, understanding and thinking still matter in programming.
Last stated 25 Aug 2026 · 3 months ago
Holds NeetCode
Similar wording
People do not value or care about something built by AI, regardless of its quality, because no personal effort went into it.
Last stated 21 Sept 2026 · 2 months ago
Holds Stefan Judis
Similar wording
If you think through the customer experience you want to create, you end up realizing hardware is a necessity.
Last stated 12 Apr 2026 · 7 months ago
Holds Evan Spiegel
Similar wording
Cheap AI-built features still cost customers complexity and cognitive load.
Last stated 2 Sept 2026 · 3 months ago
Holds Dharmesh Shah
Similar wording
A product's success depends on hardware and software working together, not hardware alone.
Last stated 10 Sept 2026 · 2 months ago
Holds M.G. Siegler
Similar wording
AI capabilities improve daily, but hardware takes months and years to ship and build at scale.
Last stated 31 Oct 2026 · 3 weeks ago
Holds Marily Nika
Similar wording
Faultless computer hardware could still support confusion, forgetting or a sense of beauty
Last stated c. 1979
Holds Douglas Hofstadter
Same subject: semiconductors
A ban on self-recursive improvement in AI models would create a big incentive for some other parts of the world to invest a lot more in compute.
Last stated 22 Sept 2026 · 2 months ago
Holds Tyler Cowen
Same subject: semiconductors
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 30 Jul 2026 · 4 months ago
Holds Jeff Dean
Same subject: semiconductors
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 26 Jul 2026 · 4 months ago
Holds Jensen Huang
Same subject: data centers
AI data centers have shifted from about 1 CPU per 8 GPUs to roughly 1:4, and the ratio could shrink to 1:1.
Last stated 24 Sept 2026 · 2 months ago
Holds Gergely Orosz
Same subject: data centers
AI workloads are best served by clouds that bundle power, data center infrastructure and GPUs and strip out unnecessary software layers.
Last stated 8 Sept 2026 · 2 months ago
Holds Chamath Palihapitiya
Same subject: data centers
Building capacity at massive scale on a single hardware generation is the trap; pacing matters more than gigawatts.
Last stated 12 Nov 2025 · a year ago
Holds Satya Nadella
Same subject: scaling laws
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 13 Mar 2026 · 8 months ago
Holds Dylan Patel
Same subject: scaling laws
Around 2030 to 2032 AI progress has to taper as compute runs out, unless very powerful capabilities arrive or fabs and investment scale faster than expected.
Last stated 8 Jul 2025 · a year ago
Holds Ryan Greenblatt
Same subject: scaling laws
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 15 Apr 2026 · 7 months ago
Holds Jensen Huang