Pinned Tweet
There’s a new AI frontier.
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
11
12
60
4,777
Jonathan Siddharth retweeted
I like Turing CEO @Jonsid's framing on how Enterprise's should think about AI over time.
For non-core workflows, its okay to rent intelligence. For important core workflows, you want to own the learning loop: have custom evals, deploy a system, record traces, fine-tune & hill climb on accuracy/cost/latency
For non-core workflows, its okay to rent intelligence. For important core workflows, you want to own the learning loop: have custom evals, deploy a system, record traces, fine-tune & hill climb on accuracy/cost/latency
8
2
28
3,296
Jonathan Siddharth retweeted
There’s a new AI frontier.
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
11
12
60
4,777
Jonathan Siddharth retweeted
The future of Evals is IRL
There’s a new AI frontier.
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
0
2
10
1,648
Jonathan Siddharth retweeted
My team at Turing builds the data that teaches models real work. Jonathan explains why deployment failures should shape that data.
There’s a new AI frontier.
It isn’t the benchmark anymore. It’s where models meet real work.
A model that tops a leaderboard doesn’t move the economy. A model that runs a month-end close does.
Models need residency. x.com/i/article/2109…
0
1
2
150
Jonathan Siddharth retweeted
Thanks Aaron Levie (Co-founder of B0X), AGI House, and Turing for a great event and panel at SFTechWeek!
Videos coming soon!
@levie
@agihouse_org
@turingcom
@Jonsid
Videos coming soon!
@levie
@agihouse_org
@turingcom
@Jonsid
2
3
22
1,516
Jonathan Siddharth retweeted
Turing CEO Jonathan Siddharth (@Jonsid) on what AI doomers are missing when they talk about AI eliminating all jobs:
“AI's biggest superpower is upleveling the type of problems humans can now solve. That's what these models help us do.”
“It's about asking the right questions & veriscope of problems that you can solve is massive.”
“AI's biggest superpower is upleveling the type of problems humans can now solve. That's what these models help us do.”
“It's about asking the right questions & veriscope of problems that you can solve is massive.”
BREAKING: The future of AI isn't open vs. closed.
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
0
1
15
1,501
Jonathan Siddharth retweeted
Turing CEO Jonathan Siddharth (@Jonsid) says AI’s bio risk is much scarier than cybersecurity risks because the downside could be catastrophic & we’d have no defense against it:
“With bio, the risk is a little bit asymmetric.”
“If somebody comes up with a crazy virus, imagine somehigh viral coefficient in terms of how quickly it spreads with a high K factor, & maybe it is much more fatal than Covid was.”
“You saw how hard it was to come up with vaccine production quickly. Deployment happens in the real world. We are limited by a lot of real world factors.”
“There is more risk in terms of how quickly we will be able to combat these.”
“With bio, the risk is a little bit asymmetric.”
“If somebody comes up with a crazy virus, imagine somehigh viral coefficient in terms of how quickly it spreads with a high K factor, & maybe it is much more fatal than Covid was.”
“You saw how hard it was to come up with vaccine production quickly. Deployment happens in the real world. We are limited by a lot of real world factors.”
“There is more risk in terms of how quickly we will be able to combat these.”
BREAKING: The future of AI isn't open vs. closed.
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
3
4
23
4,591
Jonathan Siddharth retweeted
Everything you wanted to know about AI and the future of humanity. Great discussion between @MollySOShea and @Jonsid.
BREAKING: The future of AI isn't open vs. closed.
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
1
1
8
735
Jonathan Siddharth retweeted
BREAKING: A new interview with @MollySOShea
and our CEO @Jonsid.
They discuss how AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years and break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
Watch the entire interview below.
and our CEO @Jonsid.
They discuss how AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years and break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
Watch the entire interview below.
BREAKING: The future of AI isn't open vs. closed.
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
8
6
26
7,981
Jonathan Siddharth retweeted
Turing CEO Jonathan Siddharth (@Jonsid) says there are just 3 sectors that are going to win in AI no matter what:
› Compute
› Energy
› Data
“The folks who obviously win are the folks that are inputs into the ecosystem.”
“These inputs win no matter what. You need these as you scale.”
› Compute
› Energy
› Data
“The folks who obviously win are the folks that are inputs into the ecosystem.”
“These inputs win no matter what. You need these as you scale.”
Where does money invested into the AI buildout actually go?
For every $100 flowing into the supply chain:
- $50 to chips
- $20 to power
- $15 to networking
- $15 to cooling, buildings, and land
More charts in State of Markets II: a16z.news/p/state-of-mar…
9
12
113
9,576
Jonathan Siddharth retweeted
BREAKING: The future of AI isn't open vs. closed.
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
We need a TON of both.
Turing CEO on why frontier AI & open-weight models are increasingly serving different parts of the AI stack.
“There's absolutely a place in the world for Ferraris & Koenigseggs. But there's also a place in the wontier models continue pushing the limits of intelligence. Jonathan's point isn't that one replaces the other. We need frontier AI to keep advancing capabilities & open models to diffuse those capabilities across the economy.
BTW "Only 2% of US households pay for AI" It's still early.
At the same time, AI training is moving beyond benchmarks into simulated environments built for real work, as agents move from completing tasks to operating autonomously for days, & eventually weeks, months & years.
@turingcom CEO Jonathan Siddharth (@Jonsid)
We break down what this shift means for frontier labs, enterprises & the AI stack:
› Frontier AI vs. sovereign AI
› Distillation & the shrinking model gap
› Why enterprises are moving to open models
› What AI safety & alignment teams actually do
› Reward hacking & why “if the model can figure out a way to cheat & get the reward, it will”
› Super Intelligence, recursive self-improvement & slow takeoff
𝐓𝐈𝐌𝐄𝐒𝐓𝐀𝐌𝐏𝐒
(00:00) Jonathan Siddharth, Co-Founder & CEO at Turing
(01:03) The biggest shift in AI data this year
(04:10) Why AI won't replace jobs, but uplevel them
(08:31) The arms race inside cybersecurity
(11:22) Why train AI on what it shouldn't do?
(15:45) The Hugging Face hacking incident
(22:33) How AI models cheat to win
(27:40) The AI playbook
(32:20) Open models are only 3–6 months behind
(41:30) Why chips, energy & data always win
(47:27) Is superintelligence by 2030 the goal?
(53:32) Jonathan's hottest take on AI
(55:00) The next 10 years of superintelligence
14
14
57
37,094
Jonathan Siddharth retweeted
The nation owes a debt to @jgebbia and his excellent team. They worked incredibly hard, with no compensation, to make government services work better for all Americans.
🫡 🇺🇸
🫡 🇺🇸
Imagine seeing your work announced on the national stage by the world's most gifted speakers, like @SecRubio who introduced our new design for the U.S. Passport, in collaboration with the State Department. National Design Studio is growing! ndstudio.gov/jobs
1,274
8,024
65,471
6,081,296