Transcriptly

Everyone’s Misunderstanding AI’s True Potential | Radical AI, Joseph F. Krause

EO20:43Added Aug 31, 2026

Meet Joseph Kraus, Co-Founder and CEO of Radical AI, the company trying to reinvent a 150-year-old scientific process by building scientific AGI (Artificial ...

Watch on YouTube →
Contributed by 刘嘉琪

Transcript

Transcript format

  1. Intro

  2. 00:00AI is so incredible. This technology is
  3. 00:03going to change the world. What I don't
  4. 00:04understand is why everyone is picking
  5. 00:07lowhanging fruit and solving small
  6. 00:10problems. We saw a bunch of AI companies
  7. 00:12in software or you know normal SASbased
  8. 00:16businesses. [music]
  9. 00:17Why is no one using AI to cure cancer?
  10. 00:20We are trying to disrupt the process
  11. 00:21that is 150 years old with some of the
  12. 00:24biggest companies today in materials. We
  13. 00:27want to reinvent the way we do the
  14. 00:29[music] scientific process. We don't
  15. 00:31struggle with driving one, two or
  16. 00:34[music] five percent improvements in a
  17. 00:36current system. What is incredibly
  18. 00:38challenging is novel discovery. The most
  19. 00:41important areas are the most
  20. 00:43challenging. I think those [music] are
  21. 00:45the things people should work on because
  22. 00:46if you succeed, you will fundamentally
  23. 00:49reshape human trajectory. Every single
  24. 00:52person that we hire into radical AI, we
  25. 00:55ask the same [music] question. If you
  26. 00:56are looking for a job, this is not the
  27. 00:59place for you. You're incredibly
  28. 01:01intelligent. You [music] can go get
  29. 01:02employed other places. If you are
  30. 01:04looking for a mission, then you should
  31. 01:06come work here.
  32. 01:14My name is Joseph Krauss. I'm [music]
  33. 01:16the co-founder and CEO of Radical AI. At
  34. 01:18Radical AI, we are building artificial
  35. 01:20general intelligence for scientific
  36. 01:22[music] discovery starting in material
  37. 01:25science. The material discovery process
  38. 01:27is challenging today. Processes in
  39. 01:30materials just take an incredibly long
  40. 01:32time, typically 10 plus years. Our AI
  41. 01:35agents will index millions of scientific
  42. 01:37[music] publications and learn what a
  43. 01:40field has done already. And so this
  44. 01:41process can happen [music] at an
  45. 01:43incredible speed in the comparison to a
  46. 01:46human scientist about 370 times the
  47. 01:48speed we operate as a human scientist

  48. The One Mission That Shaped My Life

  49. 01:50today.
  50. 01:54I just always was taught and really
  51. 01:56wanted to be like my dad in pursuing
  52. 01:59something that I knew I was the best in
  53. 02:01the world at and that I truly love to
  54. 02:03do. My father, he always instilled that
  55. 02:05not just about being happy about what
  56. 02:08you're doing, but much more engaged in
  57. 02:10what you're doing. Find the thing that
  58. 02:13you can be better at anyone else in the
  59. 02:15world in. Where can I make an impact?
  60. 02:17That for me has always been this guiding
  61. 02:18light into what I want to do and how I
  62. 02:20want to do it. And so when I was an
  63. 02:22undergraduate, I actually bumped into
  64. 02:24someone who was serving part-time in the
  65. 02:26National Guard. I really wanted to go to
  66. 02:29a military academy in the United States.
  67. 02:31They're incredibly prestigious. You have
  68. 02:33the opportunity to serve the country. So
  69. 02:35I thought that this would be for me. And
  70. 02:37[music] the US National Guard is a
  71. 02:39reserve component both of the Army and
  72. 02:41the Air Force. You pretty much go to
  73. 02:44drill one weekend a month and then you
  74. 02:45do two weeks in the summer of training
  75. 02:47to actually make sure you're up to code
  76. 02:49with your military training. and if you
  77. 02:50ever get deployed down range. I was
  78. 02:52like, well, this is amazing. I can
  79. 02:54continue to study, but I can also serve
  80. 02:56in the military in my free time is the
  81. 02:58way I looked at it. So, I enlisted in
  82. 03:00the military my junior year of college.

  83. The First Experience in a Strong Mission-Driven Organization

  84. 03:02Basic training is an amazing experience,
  85. 03:04right? Because you're with a bunch of
  86. 03:07people you've never met in your life.
  87. 03:08Meeting people from so many different
  88. 03:11walks of life. I met people from New
  89. 03:14York, Pennsylvania, Georgia, Maine, and
  90. 03:16the East Coast through to the Midwestern
  91. 03:19or or or western states all the way
  92. 03:20through to California coming from San
  93. 03:22Diego or Silicon Valley. And everyone
  94. 03:25was there for the same mission. They
  95. 03:27were there for the same goal of building
  96. 03:30a stronger military to protect and
  97. 03:32defend [music] the freedoms that America
  98. 03:34likes to propagate throughout the world.
  99. 03:36That was not only inspirational but was
  100. 03:38the first time I was forced to remove
  101. 03:41myself from the equation and put the
  102. 03:43mission of the team which was the
  103. 03:45battery that I was serving in before
  104. 03:47yourself. And that was an incredibly
  105. 03:49insightful lesson and so went away to
  106. 03:50basic training and then came back and
  107. 03:53finished my senior year and then went
  108. 03:54away to training again before starting
  109. 03:56graduate school at Rice.

  110. Choosing Rice to Make a Bigger Impact

  111. 03:58So for Rice there were one specific
  112. 04:01reasons. I had went to this
  113. 04:03undergraduate research symposium which I
  114. 04:05presented my work on and I won that
  115. 04:08symposium and in winning the symposium
  116. 04:10the best presentation I got to sit with
  117. 04:13a bunch of both graduate students and
  118. 04:16professors inside the program and ask
  119. 04:18them about their research and what they
  120. 04:20were working on and there was a
  121. 04:22concurrent theme through that. They were
  122. 04:24all working on problems that could have
  123. 04:25a big impact, [music]
  124. 04:27actual things that they were trying to
  125. 04:28transition at one point or another. I
  126. 04:31interviewed at a lot of graduate
  127. 04:32programs when I was thinking about where
  128. 04:34to go to school and none had felt as
  129. 04:37strong as I did when I met the faculty
  130. 04:39and students at Rice and this kind of
  131. 04:41focus on the future and what they were
  132. 04:43trying to build. So, I was doing them
  133. 04:45concurrently in graduate school and in
  134. 04:48the military at the same time, but I
  135. 04:50didn't know what I wanted to do after
  136. 04:52[music] graduate school. I didn't know
  137. 04:53if I wanted to be a professional
  138. 04:54scientist. I had thought about fields
  139. 04:56like law, patent law. So in graduate
  140. 04:59school, I did this deep dive in
  141. 05:01understanding where could I make an
  142. 05:02impact.

  143. Leaving the Research Field to Make a Bigger Impact

  144. 05:05While I was serving in the National
  145. 05:06Guard, I was also not together a
  146. 05:10scientist at the Army Research Lab. It
  147. 05:12is this corporate research lab that the
  148. 05:14US Army has a bunch of scientists
  149. 05:16working in to try to push novel research
  150. 05:20up to higher technology readiness levels
  151. 05:22or or make that science more
  152. 05:24approachable for the army. We used to
  153. 05:26have these commanders who would come
  154. 05:27through and tour the facility. We would
  155. 05:29have to tell them why the work we were
  156. 05:32doing was relevant to the army and their
  157. 05:35relevance was not the same relevance to
  158. 05:37us. They were thinking about future
  159. 05:38conflict about how to bolster the force
  160. 05:41or make it stronger. We were thinking
  161. 05:43around electrons and and hydrogen atoms
  162. 05:45and thinking about making new novel 2D
  163. 05:47materials for different technologies.
  164. 05:49And so there was this disconnect [music]
  165. 05:51between leadership at the army and the
  166. 05:53scientific perspective that we were
  167. 05:55driving. And there was this day where I
  168. 05:57had to actually explain why what we were
  169. 05:59doing in the lab could eventually impact
  170. 06:02future of warfare or future technology
  171. 06:04that the army would want to use. And
  172. 06:06that's when I realized that that is
  173. 06:08exactly what I wanted to do. I wanted to
  174. 06:11transition from this fundamental area of
  175. 06:13research driving our understanding of
  176. 06:16science to actually commercialize
  177. 06:17science. Science that could actually go
  178. 06:19into products and make a difference. And
  179. 06:22that was where things started to come
  180. 06:23together that the only way to really do
  181. 06:25this is to actually push technology in
  182. 06:28[music] a novel way in which case
  183. 06:30startups are the best in my opinion the
  184. 06:32best place to do that today. Okay. I

  185. The most important lesson from Kevin Ryan

  186. 06:34know one day I want to build a company,
  187. 06:37but I have no idea how to build it. And
  188. 06:40so I said, let me find entrepreneurs and
  189. 06:43investors who have built a company in
  190. 06:45the past and now are investing in the
  191. 06:47companies of of the future. And so I
  192. 06:49looked up a bunch of venture
  193. 06:51capitalists, entrepreneurs, and really
  194. 06:54wanted to be in New York. Actually, I I
  195. 06:55tailored my searchs to New York and I
  196. 06:57bumped into Kevin Ryan. Kevin Ryan is a
  197. 06:59prolific entrepreneur in New York City.
  198. 07:01[music]
  199. 07:01He built and ran doubleclick before
  200. 07:03taking that company public. And then he
  201. 07:04came out of that and started Ali Corp,
  202. 07:07which is both an incubation studio as
  203. 07:10well as an early stage investment firm.
  204. 07:12And when I talked to Kevin, I had told
  205. 07:14him, if you're not investing in
  206. 07:15materials, you're not going to invest in
  207. 07:16the future. And he said, "That's a big
  208. 07:18that's a big bet. Why don't you come
  209. 07:20prove that out?" And so I took a leave
  210. 07:21of absence. I moved to New York a week
  211. 07:23later, and I had a six-month internship
  212. 07:26to see if this was the field for me, if
  213. 07:28this was the career for me. Kevin had
  214. 07:30taught me early part of my career that
  215. 07:32you have to have a bias to action and
  216. 07:35just get things done. Strategy can be
  217. 07:38effective but most of the time will also
  218. 07:41lead to a dead end with no action
  219. 07:44involved. The best entrepreneurs have an
  220. 07:46immense [music] bias to action. They
  221. 07:49just get things done and in doing so
  222. 07:52watch the results unfold in front of
  223. 07:54them as they do that and then they
  224. 07:56continue to go to go through their
  225. 07:58process. [music]
  226. 07:58Startups are already hard enough and the
  227. 08:01way to make them easier is to do more
  228. 08:04than anyone in your space will do with
  229. 08:06more information than anyone in your
  230. 08:07space has with a thesis [music] that no
  231. 08:10one else has been able to come up with.
  232. 08:11That is truly what can drive good value
  233. 08:13and I think a bias to action is the most
  234. 08:15important lesson I've learned from him.

  235. Stop Using AI to Solve Tiny Problems

  236. 08:21One of my other co-founders and I were
  237. 08:23both investors at at Ali Corp and Jorge
  238. 08:26was looking into AI technology as every
  239. 08:29good investor was, but really at a
  240. 08:32fundamental level. He was reading the
  241. 08:34publications coming out on novel
  242. 08:36architectures and novel approaches to
  243. 08:39machine learning and really what the
  244. 08:41impact of the technology was. and he
  245. 08:43came over to me one day. We sat next to
  246. 08:45each other in the office and said, "You
  247. 08:47know, AI is so incredible. I am actually
  248. 08:50convinced this technology is going to
  249. 08:52change the world. What I don't
  250. 08:53understand is why everyone is picking
  251. 08:56lowhanging fruit and solving small
  252. 08:59problems. Why is no one using AI to cure
  253. 09:02cancer?" was the question he asked me.
  254. 09:04And and I said, "Well, I don't know
  255. 09:06about curing cancer, but that's a really
  256. 09:08good opportunity." And so we spent the
  257. 09:10next month and a half reading hundreds
  258. 09:13of research papers on all the different
  259. 09:15fields that we think AI could be put
  260. 09:17into. And finally it dawned on me that
  261. 09:20well material science problems with
  262. 09:22fragmentation, slowmoving ability, you
  263. 09:25know, lack of progress over short
  264. 09:26timelines. This might be a great area
  265. 09:29for AI to be implemented in. And so we
  266. 09:32took another deep dive and we read every
  267. 09:34publication we could find at the
  268. 09:36intersection of material science, AI,
  269. 09:39and as we found out, robotics. And that
  270. 09:42led us to our third co-founder, Herd
  271. 09:44Cedar, who had built an autonomous
  272. 09:47scientific lab at Lawrence Berkeley
  273. 09:49National Lab. And together with him, us
  274. 09:51three really saw and came up with this
  275. 09:54vision for where science is going to go,
  276. 09:56where AI and autonomy are going to build
  277. 09:58a new paradigm of scientific discovery,
  278. 10:01driving us from a human-driven to an AI
  279. 10:03and autonomydriven process. And we were
  280. 10:05all aligned that whether we started
  281. 10:08Radical AI or not, this is the way
  282. 10:10science was going to go. And it had to
  283. 10:13be us who were going to build this
  284. 10:15company. And so all three of us agreed
  285. 10:17and that was the formation story for
  286. 10:18radical AI. The most important areas are
  287. 10:22the most challenging. I think those are
  288. 10:24the things people should work on. I
  289. 10:26think if you look at technology today,
  290. 10:29we don't struggle with what we call
  291. 10:32optimization problems at Radical AI,
  292. 10:34driving one, two, or 5% improvements in
  293. 10:37a current system. We're actually quite
  294. 10:39good at that. What is incredibly
  295. 10:41challenging is novel discovery. And
  296. 10:44that's where most hard problems lie. And
  297. 10:46when us three wanted to form the company
  298. 10:48and we're going to build radical AI, we
  299. 10:51had this opinion that if you want to
  300. 10:52impact the most important industries in
  301. 10:55the world, automotive and aerospace,
  302. 10:57manufacturing and defense, climate
  303. 10:59energy, semiconductors, all of them are
  304. 11:01a direct result from materials R&D. But
  305. 11:04these processes in materials just take
  306. 11:06an incredibly long time, typically 10
  307. 11:08plus years and an exorbitant amount of
  308. 11:11cost to go from a novel discovery to a
  309. 11:13scale material system. And this was the
  310. 11:15exact place [music] that we could make a
  311. 11:17massive impact. We think AI and autonomy
  312. 11:20[music] are going to drive change in the
  313. 11:22process and in doing so can actually
  314. 11:25unlock and remove materials as our
  315. 11:27biggest barrier to some of our most
  316. 11:29important industries. And I think if you
  317. 11:30are aligned on that mission, then the
  318. 11:34problems that you want to solve are
  319. 11:35naturally going to be hard because if
  320. 11:37you succeed, you will [music]
  321. 11:40fundamentally reshape human trajectory.
  322. 11:42And that is an incredibly impactful
  323. 11:44thing to be able to do and also
  324. 11:46incredibly important to the future of
  325. 11:48the world.
  326. 11:50I'm Taylor Rottwell, the founder and CEO
  327. 11:52of Laravel. Our mission is to help
  328. 11:55developers build, deploy, and monitor
  329. 11:57applications consistently and
  330. 11:58beautifully.
  331. 12:00Starting projects is great, but more
  332. 12:02importantly, you need to ship them.
  333. 12:03Great ideas mean nothing until they
  334. 12:05ship.
  335. 12:07That's exactly why I built Laravel, to
  336. 12:09bring your entire journey from idea to
  337. 12:11production into one seamless flow.
  338. 12:14It's more than a PHP framework. It's a
  339. 12:16complete connected ecosystem built for
  340. 12:18speed and confidence. We provide
  341. 12:21elegant, opinionated solutions for
  342. 12:23everything your modern web app needs.
  343. 12:25And you can deploy in under a minute
  344. 12:27with Laravel Cloud, straight from Git.
  345. 12:29Zero config needed. Laravel now powers
  346. 12:32nearly 600,000 live websites, helping
  347. 12:35developers worldwide ship their ideas
  348. 12:37faster.
  349. 12:39Build with elegance. Deploy with
  350. 12:41confidence. Monitor with precision.
  351. 12:44Stop configuring. Start shipping. Visit
  352. 12:46larav.com today.

  353. How I Raised $55M in 45mins

  354. 12:56We're going to leave Ali Corp and we're
  355. 12:58going to raise money of course to start
  356. 13:00the company and Ali Corp as I mentioned
  357. 13:02likes to incubate companies and we were
  358. 13:03incubating this company inside the Ali
  359. 13:05Corp umbrella. So we knew Kevin had told
  360. 13:07us they are going to put some money in
  361. 13:09in this incubation setting and then
  362. 13:11we'll roll out from there and we knew we
  363. 13:13needed a lot of capital. Material
  364. 13:14science is a hard business to build in
  365. 13:16and we are building a full stack
  366. 13:18solution which requires a lot of
  367. 13:20capital. So we knew we wanted to raise a
  368. 13:22fairly large preede round and so we went
  369. 13:25to Kevin and said look we're going to
  370. 13:26we're going to start this company put
  371. 13:27together a whole 100page pitch deck on
  372. 13:29why we think the opportunity is now
  373. 13:31where the technology is today why
  374. 13:33interdisiplinary approach is incredibly
  375. 13:35important to solving the problem. The
  376. 13:37reason AI is so impactful in this space
  377. 13:40is its ability to index information. The
  378. 13:43best example that we always like to give
  379. 13:45is Alph Go. [music] It was this AI that
  380. 13:48was built by the Google team and it
  381. 13:50played the best player in the world, a
  382. 13:51game of Go. And the best part of the
  383. 13:52story is this infamous move it makes
  384. 13:54where everyone watching thinks the AI
  385. 13:56makes a mistake and in reality it has
  386. 13:58indexed so many games of Go. It made a
  387. 14:00move the human brain has never thought
  388. 14:01of before. [music] And when we take that
  389. 14:03to science, all of our limitations in
  390. 14:06science come from the human brain. How
  391. 14:08many papers can we read? How many things
  392. 14:11can we simulate? [music]
  393. 14:12how many experiments can we run and then
  394. 14:14how do we tie all those together to make
  395. 14:16a new hypothesis on what we want to
  396. 14:18build. But if you bring AI and autonomy
  397. 14:20to the center of that, you unlock this
  398. 14:23indexing problem that a human scientist
  399. 14:25feels. You can read millions of
  400. 14:27publications, simulate billions of
  401. 14:30materials and test thousands of them
  402. 14:33connecting all this information in real
  403. 14:35time. And so if you can do that, the
  404. 14:39world that we think we can really build
  405. 14:41is one that is what we call inversely
  406. 14:44designed where we are no longer making
  407. 14:47materials and then looking for a
  408. 14:49solution to solve. We are actually
  409. 14:52taking our hardest problems and
  410. 14:54inversely designing material from that.
  411. 14:56And human scientists can't do that
  412. 14:59[music] today or if they can can do so
  413. 15:03in only very very long time frames
  414. 15:06whereas AI and autonomy can do that
  415. 15:09incredibly fast and incredibly [music]
  416. 15:12efficient. What we actually think AI can
  417. 15:14do is not replace that scientist but
  418. 15:17rather give them the tools and the
  419. 15:18capabilities [music] to index across a
  420. 15:21multitude of different experimental
  421. 15:22processes. So they are spending their
  422. 15:25time thinking around what next material
  423. 15:27to build or what next problem to solve
  424. 15:30and not on going through the monotonous
  425. 15:32process of [music]
  426. 15:33fundamental research. That is really
  427. 15:35where we see the impact on AI for
  428. 15:37science and how we are building the
  429. 15:40future of the scientific process. And to
  430. 15:42Kevin's credit he said you're not going
  431. 15:44to raise money anywhere else. I want to
  432. 15:47give you all of it. And so he was the
  433. 15:48only investor. We raised our preede
  434. 15:50round in about 45 minutes uh which was
  435. 15:52great. and we were able to get started
  436. 15:54and start building and us three started
  437. 15:56recruiting from that time on and built
  438. 15:58the company that that we have today.

  439. The 51% Rule

  440. 16:06We are incredibly
  441. 16:08passionate about the culture that we
  442. 16:10build at the company. Every single
  443. 16:11person that we hire into Radical AI, we
  444. 16:14ask the same question. If you are
  445. 16:16looking for a job, this is not the place
  446. 16:18for you. You're incredibly intelligent.
  447. 16:20You can go get employed other places. If
  448. 16:22you are looking for a mission then you
  449. 16:24should come work here. Every person that
  450. 16:27comes to radical AI deeply deeply
  451. 16:30believes in the mission that we are
  452. 16:31going after. [music] And the way that we
  453. 16:33keep that alignment is through culture.
  454. 16:36We are never afraid to fail. We actually
  455. 16:40know failure [music] is a part of the
  456. 16:43discovery and learning process. And so
  457. 16:45we will push aggressively, relentlessly
  458. 16:49to drive technology to rethink from
  459. 16:52first principles and to build a
  460. 16:54connected system that can truly discover
  461. 16:56novel materials. So we adopted a rule
  462. 16:58very in the company. I believe SpaceX
  463. 17:00started it called the 51% rule where
  464. 17:03when you are at 51% confidence on a
  465. 17:06decision, you just make that decision.
  466. 17:08And the reason why is you will actually
  467. 17:11cause more time and less efficacy across
  468. 17:14the company by debating on decisions
  469. 17:17that you could otherwise quickly decide
  470. 17:19on and see what the result of that
  471. 17:20decision is going to be. When it comes
  472. 17:22to 51% two different things that we
  473. 17:24think about first how big is the
  474. 17:26decision and second what are the risks
  475. 17:29if the decision is wrong and those two
  476. 17:32together allow us to actually identify
  477. 17:34when we are at 51% or not. Of course, we
  478. 17:37do a bunch of research and have a bunch
  479. 17:40of conversations and debate around
  480. 17:42getting to that confidence interval of
  481. 17:4451%. But when it is a decision in
  482. 17:47dayto-day, for example, it's not year
  483. 17:49defining or or really decade defining
  484. 17:52and what we're trying to go after. Well,
  485. 17:54then [music] we can have confidence to
  486. 17:56make it quickly. And when we think about
  487. 17:58what the other side of that decision is,
  488. 18:00what is the worst case scenario that
  489. 18:02happens that usually alerts you to you
  490. 18:05are already at 51% or you are nowhere
  491. 18:08near 51%. And that is a great checkpoint
  492. 18:11that we ask ourselves and everyone in
  493. 18:13the company and constantly pushing
  494. 18:15towards uh getting getting at quick
  495. 18:17decision-m but effective decision-m as
  496. 18:19well. You know, 51% is not about just
  497. 18:22making decisions fast for fast sake. 51%
  498. 18:26is making decisions that you are already
  499. 18:29going to make, just making them earlier
  500. 18:31and then dealing with the effect of not
  501. 18:34always being 100% correct. Because no
  502. 18:36matter if you think about a decision for
  503. 18:372 weeks or 2 years, there's still a
  504. 18:40likelihood that you're not going to be
  505. 18:41correct. And because we operate the
  506. 18:43company that way, failure is inherent.
  507. 18:46You will automatically fail if you come
  508. 18:48work at Radical AI in something that you
  509. 18:50do. We don't view someone as failing or
  510. 18:54a project as failing. We view it as
  511. 18:56learning and an opportunity to recreate
  512. 18:59the process that we just thought that we
  513. 19:01should build. And so for us, I actually
  514. 19:03think we go deeper into the essence of
  515. 19:06asking why and first principles. And we
  516. 19:08don't view things as failure. It is just
  517. 19:10normal to try things that don't work,
  518. 19:13learn from them, and try again at

  519. The One Piece of Advice to My Past Self

  520. 19:15Radical AI. One of my favorite quotes in
  521. 19:18the world is, you know, Steve Jobs is
  522. 19:19connecting the dots. You can never
  523. 19:21connect the dots looking forward. You
  524. 19:23can only connect them looking backwards.
  525. 19:25This is an exact moment where I didn't
  526. 19:26know where this dot was going to lead.
  527. 19:28But now when I look back, it was
  528. 19:30imperative that I went through that
  529. 19:32experience and [music]
  530. 19:33lived through the frustrations of
  531. 19:35materials research today. You need dots
  532. 19:37to be able to connect. And while you
  533. 19:39might not know where this dot is going
  534. 19:41to make an impact, you never know where
  535. 19:43it is going to [music]
  536. 19:44or how it is going to impact what comes
  537. 19:47next. And so I didn't care that I didn't
  538. 19:49know what was coming next.
  539. 19:50[music]
  540. 19:50I didn't care if it was going to work or
  541. 19:52not. I didn't care if I was going to be
  542. 19:55a scientist or not in in the real
  543. 19:57retrospect. What I cared about was
  544. 19:59continuing to push forward and searching
  545. 20:01for the answer. And I think that's truly
  546. 20:03how I ended up getting to the answer.
  547. 20:04So, I just reminded myself every day
  548. 20:06that this will lead to something. You
  549. 20:08have to never ever give up no matter the
  550. 20:11scenario. We want to remove materials as
  551. 20:14the blocker to those innovations. And
  552. 20:16that's why we don't think there's an end
  553. 20:18[music] to the company. We think this
  554. 20:19company will be 100, 200 years old. It
  555. 20:22will way outlive myself and and the two
  556. 20:24co-founders. This is a company and a
  557. 20:26technology [music] that when it succeeds
  558. 20:29will truly be able to create a world
  559. 20:32[music] not today thought possible.