Principal Robotics + AI @amazon | building physical AI in the real world | MSc AI @LJMU | founder @ghostcentauri | prev @framebridge, @teespring (YC)
Kentucky
Joined December 2011
- Tweets4.6K
- Following4.8K
- Followers1.6K
- Likes37.8K
Pinned Tweet
Worked on Eluna for over a year here, and while we're not always able to talk about the work… stoked more of it is coming public!
aboutamazon.com/news/operati…
Tyler Wolfe retweeted
Kentucky to be the 36th US state to build a grid-connected nuclear plant,
and the first new state to do so, at GW-scale, in over 50 years!
1 - Pennsylvania - Shippingport - 1954
2 - New York - Indian Point - 1956
3 - Illinois - Dresden - 1956
4 - Michigan - Fermi - 1956
5 - Massachusetts - Yankee Rowe - 1957
6 - California - Humboldt Bay - 1960
7 - Wisconsin - LaCrosse - 1963
8 - Connecticut - Haddam Neck - 1964
9 - New Jersey - Oyster Creek - 1964
10 - South Carolina - Robinson - 1967
11 - Florida - Turkey Point - 1967
12 - Alabama - Browns Ferry - 1967
13 - Minnesota - Monticello - 1967
14 - Vermont - Vermont Yankee - 1967
15 - Nebraska - Cooper - 1968
16 - Maryland - Calvert Cliffs - 1968
17 - Virginia - Surry - 1968
18 - Colorado - Fort St. Vrain - 1968
19 - Georgia - Hatch - 1968
20 - Arkansas - ANO - 1968
21 - Maine - Maine Yankee - 1968
22 - Oregon - Trojan - 1970
23 - North Carolina - Brunswick - 1970
24 - Iowa - Duane Arnold - 1970
25 - Tennessee - Sequoyah - 1970
26 - Ohio - Davis Besse - 1970
27 - Washington - Columbia - 1972
28 - Mississippi - Grand Gulf - 1974
29 - Louisiana - Waterford - 1974
30 - Texas - Comanche Peak - 1974
31 - Missouri - Callaway - 1975
32 - Arizona - Palo Verde - 1976
33 - New Hampshire - Seabrook - 1976
34 - Kansas - Wolf Creek - 1977
...
35 - Wyoming - Kemmerer - 2026 (TerraPower Natrium)
...
36 - Kentucky - Prosperity One - 2028 (Westinghouse AP1000)
Who's next?
Alaska, Delaware, Hawaii, Idaho, Indiana, Montana, Nevada, New Mexico, North Dakota, Oklahoma, Rhode Island, South Dakota, Utah, West Virginia
Tyler Wolfe retweeted
A proposed supersite of AP1000s has just been announced in Appalachia.
The Nuclear Company just revealed plans to build up to six Westinghouse AP1000s on a single site in Lewis County, Kentucky.
If all six reactors get built, it would be the largest nuclear site ever developed in the U.S.
If the reactors are built in pairs, this would be the closest America has come to copying China’s recipe for nuclear success.
My team and I have spent the past year studying China’s nuclear program through Chinese primary sources, and there’s a clear pattern to their builds:
- 6 reactors planned per site built two at a time;
- Using a standard design they repeat over and over;
- With the same crews moving from one pair to the next.
Although a frozen reactor design is nice, nothing gets you more standardization than building on the same spot with the same team.
The site appears to sit along a major 765 kV transmission line in the Ohio River corridor (not a coincidence, I'm sure). But a full 6.6 GW will almost certainly need grid upgrades including additional outbound transmission.
This is a proposed project, not a done deal. And we're far enough out from Vogtle that building these reactors will require significant workforce training, standing up of supply chains, and coordination across many parties.
But oh boy, would be massive if The Nuclear Company pulls this off.
Tyler Wolfe retweeted
The Nuclear Company plans to deploy a series of gigawatt-scale campuses with fleets of reactors, totaling more than 40 GW. The first step in that journey was announced this morning with Prosperity One in Lewis County, Kentucky
Tyler Wolfe retweeted
Robot intelligence is becoming cheap and abundant!!
We built a robotics data engine with Astra-level performance at 125x lower cost. And we are open-sourcing the weights today! Introducing Flora, a suite of open models to parse the physical world and turn raw footage into rich training signals.
Flora already powers our in-house data engine, labeling 200K hours of videos a day on 500 GPUs. We created it to scale the frontier for robot learning. Now you own the weights too.
Flora gives you:
- High-quality, timestamped captions to curate and filter human egocentric and robot data;
- Dense progress rewards for quality control and RL post-training;
- Rubrics scoring for real-world benchmarking
Try it today, links in thread:
Tyler Wolfe retweeted
BREAKING: @noxmetals, the YC-backed manufacturing startup, says it has reached a $25M annualized revenue run rate in less than a year.
The Detroit-based startup is automating metal cutting and supply for the American industrial base.
ALT Source: Nox Metals
Really exciting to see SMRs bringing new life to former coal sites!
Amazon and X-energy have inked another deal to deploy nuclear power to the grid. This one will generate nearly 1 gigawatt of electricity.
spr.ly/6019BHrMp3
Love to see it!
US utility American Electric Power has partnered with Amazon and X-energy to evaluate a potential 960 MW project - featuring a dozen Xe-100 SMRs - at a former coal plant site in Sullivan County, Indiana, in the US.
world-nuclear-news.org/artic…
Tyler Wolfe retweeted
This report estimates the U.S. AI lead over China at just six months.
It cannot be overstated how vital it is that we prevent reactionary, poorly written regulations that would smother American innovation.
And the critical compute infrastructure behind it all must be built.
Big new IFP report from @GeorgiaCAdamson, @KonstantinPilz, and me on the US-China AI competition.
We spent the last 6 months digging into how Chinese AI companies’ access to compute and distillation of American models translate into the current US-China capabilities gap.
The result is a 23k-word research report, including an interactive analytical model that visualizes changes in the US AI lead under different policy interventions.
Our findings:
1. The US AI lead over China is currently only around 6 months.
2. The US could roughly double this lead, to over a year, using authorities already available to the federal government.
3. On chips: China’s access to AI chips will roughly double over the next year, with chips housed in data centers overseas accounting for roughly half of China’s 2027 compute.
4. On distillation: the current ~16x US training compute advantage would normally yield a 13-month lead. But we estimate adversarial distillation halves this to around 6.5 months.
5. Increasing the lead would have some obvious benefits:
- Give the US broadly a greater cyberdefensive advantage against Chinese threat actors.
- Give US AI companies breathing room to spend more R&D resources on making their systems safer and more secure, without fear of being overtaken by less safety-conscious Chinese competitors.
6. Provide the US government more leverage in negotiations with China about the future of AI.
7. The two most important interventions are:
- For BIS: Create a new end-use control to restrict China’s remote access to training clusters abroad — China’s largest and fastest-growing source of frontier-scale compute.
- For DOJ/FTC: Create a narrow safe harbor to allow US companies to coordinate on countering Chinese adversarial distillation.
8. Both interventions (and all the others we discuss in the report) are things the federal government could do today.
9. To estimate the effects of each policy intervention, we model the US lead as a function of its compute advantage over China and China’s ability to adversarially distill the capabilities of the US frontier.
10. The model makes many simplifying assumptions, but provides a useful way to assess and rank the impact of different policy interventions at a high level.
Read the full report here: ifp.org/how-to-increase-amer…
Tyler Wolfe retweeted
Kind of disappointed that none of the frontier models refuse by saying "I'm sorry Dave, I'm afraid I can't do that."
