agentic jensen intelligence

agenticai.docx 

(more work needed - who's in dc arm office of jensen ceo prodigee rene .). doznes of engineers who have been at nvidia for 25 years doing life's work ...

notes on contents of Jensen guide to ai collaborations with productive human brainpower- file by year/month with additions to those in above doc labeled XX ;eg xx 2007.1..Jensen Huang interview with Taiwan's Maurice Chang (helps explain how much AI20s depends on Taiwanese people  https://www.youtube.com/watch?v=u-x7PdnvCyI

-references to 25X involve data sovereignty applications deploying 25 times energy more efficient supercompute  ..xx 24.9.1 https://www.youtube.com/watch?v=doJDuLMnaWc  jensen testimony to dc bipartisan committee - launch of 25x supercompute - ability to build one every 3 months with sovereign data regional partners national energy week

25.6.1 Macron (3rd leg king charles ai nov 2023 summit- modi to be 4th) Lecun Mistral Mensch- Macron invites top deep 10 french corporates eg schneider volia ... to design deep search llms and startup infrastructure with Mistral (meta fair llama 3 with Yann Lecun), edge ai (eg dell),- first time ever EU country supported youth ai startups/  XX25.6.2 jensen uk ai summit with PM Sharmer & 10 corporates; keynote-

typical q to grok june 2025- what intelligence uses and data sovereignty maps will be priritised by places with 25x supercomputers

REINTRODUCTION TO JENSEN HUANG VIEWING

His tips are often brutally honest; unconventional wisdom has its place in studying entrepreneurship- at least i feel so ,my dad made this argument in The Economist from 1976

2013.1 how nvidia started in 1993 

 XX25.6.2 https://www.youtube.com/watch?v=SsNS3Xm9ig4 london tech week ai summit jensen keynote included maps of euro countries nvidia open partners (bloomberg coverage hunag uk prime minister Starmer 

25.5.1 update taiwan 25x supercompute partners inluding 8 deepest corporates of taiwan as deep ai nation since 1987 national startegy of changs ai foundry and foudational maps of sme value chains of ai japan-korea--taiwan-Guo , H Li - (coming grok assessment of main applications of taiwan 25x compute

xx25.4.1 https://www.c-span.org/person/jensen-huang/141815/jensen and leading global ai industry friends collect solutions for and with president trump

xx 25.4.2 stephen witt author of thinking machines reviews first 25 years of jensen's 21st C https://www.c-span.org/program/book-tv/after-words-with-stephen-wit... 1 minute of jensens time is worth a million $ approx- at same time deep family kids now grown up jensen kids 28.30 come from louis vuitton and cocktail bar taiwan before now at nvidila- see also 2010.1 family launch stanford's deep learning lab- hugely supportive engineering wife lori- noth enegineers at orgeon state (big gift)                                                                                                       jargon check  min 5.5what is a gpu?  14.30 what is supercomputer's operating language  "platform" cuda (and how did this emerge from lifework inspirations inspirations like medical imaging, early material after big bang and weather forecasting ) 18 mins headhunt bill daly 2009 (science users of supercompute before ai) - what if i can help your science work be 1000 times faster (nvidia enters mamograms early 200s as well as coding pixels with steve jobs pixar) 44.00 what is omniverse - eg robot simulation dishwashing by breaking a million dishes in virtual world before. 48.30 pivoting whole cokpany on neural nets before even the ai expsrts understood!.. (2009 stnafird deep learning lab)

25.3.1 gtc 2025 keynote- often jensens main annual update in usa. on eoof big abbouncement is Netwon Robot lab as partnership between nvidia (deep mind hassabis and google) and diseny now owners of pixar. This brings full circle one of earliest pixel coding partnerships between nvidia and pixar then owned by steve jobs. Newton Lab is for robots to train on physics of touch 

25.3.2 At GTC jensen shares the stage with quantum companies to clarify their visions

25.1.1 At Consumer Electronics Show Jensen focused on advancing AI across industries, particularly through physical AI (AI that perceives, reasons, and acts in the real world) and NVIDIA’s new products. The keynote emphasized gaming, robotics, autonomous vehicles (AVs), and personal AI computing, powered by NVIDIA’s Blackwell architecture and Cosmos platform (this platform - an open-source world foundation model (WFM) platform  (on Github & Hugging Face) designed to train and deploy physical AI for robotics and AVs

24.11.1 In Tokyo Jensen Huang interviews Softbank Maya Son. Celebrates builinding one of world's first 25 time X supercomputes in Japan around Maya Son's frineds. Back before covid Maya Son owned Arm and valued Nvidia. Coviud firced sell off his most profiatbale investment ie both Arm and Nvida. But ceos on siftbank, nvidia, arm have mainatied same vision - supercomputes need best of arm's cpus as well as nvidia's gpus.  (If it hadnt been for Biden Justice departments mono0oly decision, nvidia and arm would be one company - see XX 24.9.1 https://www.youtube.com/watch?v=g5llbNt7_Ik

24.11.2 At CS24 annual (annual summit for high performsamnce computing developers) Buck and Jensen announce cuPyNumeric - this simplifies connectiosn python science programmers need between up to 450 python libraies nvidia offers

24.3.2 jensen brings together the 7 who broke through with transgormer model attemtion is all you need

24.3.1 Digital Twins was one of the interesting foci of jensens gtu 2024 keynote. Increasingly if you are building a factory  for soace with robots it makes sense to digitally architect this first. That way you can uoyodate or replicate design of human and robot interaction. Notably Jensen and Taiwan companies value digital twins in case they are asked to move supply chains etc. (taiwan has over 40 years of being world class leader at advanced manufacturing dssign - clarifying exactly how to move designs with hundreds of parts has become too detailed a digital infrastructure game not to use AI.

24.3.3 GTc24 sees updates to Jensen's t=vision to 3 core models in one needed for erath 2.0 not just to be best weather forecast model but to plug and plau various climate chamge games. Jensen has announced with Taiwan the intention to make Taiwan world chlass epicentre of earth 2.0 modeling- Taiwan's national investment arm and data from IBM's weather company are early partners. Both Taiwan and Jaoan currently have to import most of their energy needs- suggesting they have win-win needs to understand future of energy

xx24.11.1 Indonesia AI Day https://www.youtube.com/watch?v=w64JT0HwqHI  with Jensen Huang and Accenture's Julie Sweet  https://indonesiaaiday.com/#section-speakers  - see also numbers of accenture staff trained in AI

Load Previous Replies
  • up

    chris macrae

    “A Global Perspective on the Drug Discovery Ecosystem”

    Kenichiro Watanabe, Director, National Healthcare Policy Secretariat, Cabinet Office
    Rami Suzuki, Representative Director CEO, ARC Therapies Inc.
    Shinichiro Fuse, TPG Life Sciences Innovations / Partner and Managing Director
    Dan Kemp, CEO, Shinobi Therapeutics, Inc.
    Mikkel Skovborg, Senior Vice President, Innovation, Novo Nordisk Foundation,

    <Moderator>
    Akihiko Soyama, CEO, Life Science Innovation Network Japan / Specially Appointed Professor, Tohoku University

    10:50~11:00Break

    11:00~11:50Special Session 1 Supported by Novartis Pharma KK

    “Fueling the Future of Innovation”

    John Paul Pullicino, President and Representative Director, Novartis Pharma KK
    Fumiaki Ikeno, Researcher, Stanford University
    Kazumasa Oguro, Professor, Faculty of Economics, Hosei University
    Masato Iwasaki, Senior Executive Fellow, IGPI Group, Inc.

    <Moderator>
    Yasuko Shoji, Manager, Research Unit / Medical & Healthcare Institute, Nikkei BP Intelligence Group, Nikkei Business Publications, Inc.

    11:50~13:30Panel Session 2

    “Optimization of Institutional Design to Support Drug Discovery” (tentative)

    Tadayoshi Mizutani,Director, Policy Planning Division for Pharmaceutical Industry Promotion and Medical Information Management, Health Policy Bureau, Ministry of Health, Labor and Welfare
    Masaaki Miyakawa, Executive Board Member, Japan Medical Association
    Hitoshi Kuboniwa, Japan Bioindustry Association,Chairman Steering Committee
    Yukiko Nishimura, NPO ASrid, President
    Yasuhiro Fujiwara, Chief Executive, Pharmaceuticals and Medical Devices Agency
    Yuji Kashitani, Executive Director, Global Regulatory Policy & Innovation Japan, Takeda Development Center Japan
    Masanobu Saito, Corporate Officer, Head of Value & Access, Japan, Novartis Pharma KK
    Masahisa Jinushi, Executive Direcotr, Head of Medical Affairs, Gilead Sciences KK

    <Moderator>
    Shintaro Sengoku, Professor, School of Environment and Society, Institute of Science Tokyo

    13:30~13:40Break

    13:40~15:00Panel Session 3

    “Challenges in fostering drug discovery startups from Japan: Creating a Virtuous Circle”

    Hirokazu Shimoda, Director, Bio-Industry Division, Commerce and Service Industry Policy Group, Ministry of Economy, Trade and Industry
    Yoshiki Sawa, President,Organization of Future Medicine 
    Toshio Fujimoto, Chief Executive Officer, iPark Institute Co., Ltd.
    Yusaku Katada, Restore Vision Inc. CEO
    Morten Sogaard, Head, Innovation Lab, Astellas Pharma
    Hiroo Igarashi, President & Representative Director, Pfizer Japan Inc.

    <Moderator>
    Masamitsu Harata, Chairman, CEO and Founder at Human Life CORD Japan Inc.

    15:00~15:50Special Session 2 Supported by Novo Nordisk Pharma Ltd.

    “Challenges and Value Assessment of Obesity: Exploring the Ecosystem”

    Koutaro Yokote, President, Japan Society for the Study of Obesity/ President, Chiba University
    Ataru Igarashi, Associate Professor, Dept. of Public Health and Health Policy, Graduate School of Pharmaceutical Sciences, The University of Tokyo

    15:50~16:00Special Speech 1

    Fumio Kishida, Member of the House of Representatives

    16:00~16:50Special Session 3 Supported by Gilead Sciences, Inc.

    Gaku Hashimoto, Professor, Faculty of Health and Welfare Services Administration, Kawasaki University of Medical Welfare (Former Member of the House of Representatives)
    Wataru Sugiura, Director General, Center for Clinical Science, Japan Institute for Health Security
    Hiroto Araki, Director, Infectious Diseases Control Division, Ministry of Health, Labor and Welfare
    Toshio Fujimoto, Chief Executive Officer, iPark Institute Co., Ltd.
    Kennet Brysting, President and Representative Director of Gilead Sciences KK


    Toshihiko Takeda, Senior Advisor, Boston Consulting Group

    16:50~17:00Break

    17:00~18:00NIKKEI Drug Discovery Startup Pitch
    Finalist Presentation DAY1

    <Commentator>
    Masamitsu Harata, Chairman, CEO and Founder at Human Life CORD Japan Inc.
    Toshio Fujimoto, Chief Executive Officer, iPark Institute Co., Ltd. 
    Tsuyoshi Tsujimura, Investment Principal, Investment Department, Kyoto University Innovation Capital Co., Ltd.
    Masakazu Komahashi, General Manager, Investment Department Ⅲ, SMBC VENTURE CAPITAL CO., LTD.

    【Day2】 Wednesday, June 25

    9:00~9:20Special Speech 2

    Akihisa Shiozaki, Member of the House of Representatives

    9:20~10:40Panel Session 4

    “Drug Discovery Frontiers from Japan: Connecting with the Global Community”

    Kanae Kurata, Director, Life Sciences Division, Research Promotion Bureau, Ministry of Education, Culture, Sports, Science and Technology (MEXT)
    Shin Kaneko, Professor, Center for iPS Cell and Research Application, Kyoto University
    Professor, Faculty of Medicine, University of Tsukuba
    Kiyohumi Kaneshiro, Chief Financial Officer, PeptiDream Inc.
    Takeyuki Akiyama, Director, Japan Alliance for Lysosomal Disease Patient Organizations
    Shinichiro Komoto, Eight Roads Ventures Japan・Partner
    Hiroshi Miyake, Chief Executive Officer, Chordia Therapeutics

    <Moderator>
    Aya Kubota, Editor in Chief, Nikkei Biotechnology & Business, Nikkei Business Publications, Inc.

    10:40~11:30Special Session 4 Supported by Japan Pharmaceutical Manufacturers Association (JPMA)

    “Co-creation and Innovation from the Patient's Perspective”

    Jin Shiomura, Founder, Managing Director & CEO of Nobelpharma
    Sumito Nishidate, Chairman of the GIST & Sarcoma Patients and Families Association “NPO GISTERS”
    Kazuhiko Mori, Senior Managing Director of Japan Pharmaceutical Manufacturers Association
    Masami Sakoi, Chief Medical and Global Health Officer of Ministry of Health, Labor and Welfare

    <Moderator>
    Asuka Miyabashira – President of Japan Pharmaceutical Manufacturers Association

    11:30~11:40Break

    11:40~12:10Special Speech 3

    Kazuto Ihara, Vice-Minister of Health, Labor and Welfare

    12:10~13:30Panel Session 5

    “Innovation in Drug Discovery Driven by AI and DX”

    Yasushi Okuno, Professor, Department of Biomedical Data Intelligence, Graduate School of Medicine, Kyoto University
    Akira Izumi, CEO & Founder, RyuWell Co.,LTD
    Hidenobu Ishizaki, Executive director, Center for Development of Advanced Cancer Therapy, Cancer Institute Hospital, Japanese Foundation for Cancer Research (JFCR)
    Hiroyoshi Toyoshiba, Director/ Chief Technology Officer
    Hiroyuki Tsunoda, Deputy Head of Research Division, Chugai Pharmaceutical Co., Ltd.

    <moderator>
    Masanori Shindo, Deputy General Manager, Life Intelligence Consortium general incorporated association</moderator>

    13:30~13:40Break

    13:40~14:00Special Session 5 Supported by Johnson & Johnson Innovative Medicine

    “Leading where medicine is going”

    Sarah Brennan, Company Group Chairman, Global Commercial Strategy Organization, for Johnson & Johnson Innovative Medicine

    14:00~15:30Panel Session 6

    “Toward a Japanese Drug Discovery Ecosystem”

    Shintaro Sengoku, Professor, School of Environment and Society, Institute of Science Tokyo
    Asuka Miyabashira, President of Japan Pharmaceutical Manufacturers Association
    Takahiko Iwaya, Chair of European Federation of Pharmaceutical Industries and Associations, Japan (EFPIA Japan)
    Hans Klemm, Japan Representative, Pharmaceutical Research and Manufacturers of America (PhRMA)
    Hitoshi Nakagama, Japan Agency for Medical Research and Development/President
    Kazuto Yamada, President and CEO, Japan Tissue Engineering Co., Ltd.

    <Moderator>
    Kiyoshi Ando, ​​Senior Staff Writer, Nikkei Inc.

    15:30~15:40Break

    15:40~16:40NIKKEI Drug Discovery Startup Pitch
    Finalist Presentation DAY2

    <Commentator>
    Masamitsu Harata, Chairman, CEO and Founder at Human Life CORD Japan Inc.
    Toshio Fujimoto, Chief Executive Officer, iPark Institute Co., Ltd. 
    Tsuyoshi Tsujimura, Investment Principal, Investment Department, Kyoto University Innovation Capital Co., Ltd.
    Masakazu Komahashi, General Manager, Investment Department Ⅲ, SMBC VENTURE CAPITAL CO., LTD.
    *Please note that the program and contents of the presentations may be changed without notice.


    Speaker

    Fumio Kishida

    Member of the House of Representatives

    Hideki Murai

    Member of the House of Representatives

    Akihisa Shiozaki

    Member of the House of Representatives

    Kazuto Ihara

    Vice-Minister of Health, Labor and Welfare

    Kenichiro Watanabe

    Director, National Healthcare Policy Secretariat, Cabinet Office

    Rami Suzuki

    Representative Director CEO, ARC Therapies Inc.

    Shinichiro Fuse

    TPG Life Sciences Innovations
    Partner and Managing Director

    Dan Kemp

    CEO, Shinobi Therapeutics, Inc.

    Mikkel Skovborg

    Senior Vice President, Innovation, Novo Nordisk Foundation,

    Akihiko Soyama

    CEO, Life Science Innovation Network Japan / Specially Appointed Professor, Tohoku University

    John Paul Pullicino

    President and Representative Director, Novartis Pharma KK

    Fumiaki Ikeno

    Researcher, Stanford University

    Kazumasa Oguro

    Professor, Faculty of Economics, Hosei University

    Masato Iwasaki

    Senior Executive Fellow, IGPI Group, Inc.

    Yasuko Shoji

    Manager, Research Unit
    Medical & Healthcare Institute, Nikkei BP Intelligence Group, Nikkei Business Publications, Inc.

    Tadayoshi Mizutani

    Director, Policy Planning Division for Pharmaceutical Industry Promotion and Medical Information Management, Health Policy Bureau, Ministry of Health, Labor and Welfare

    Masaaki Miyakawa

    Executive Board Member, Japan Medical Association

    Hitoshi Kuboniwa

    Japan Bioindustry Association,Chairman Steering Committee

    Yukiko Nishimura

    NPO ASrid, President

    Yasuhiro Fujiwara

    Chief Executive, Pharmaceuticals and Medical Devices Agency

    Yuji Kashitani

    Executive Director, Global Regulatory Policy & Innovation Japan, Takeda Development Center Japan

    Masanobu Saito

    Corporate Officer, Head of Value & Access, Japan, Novartis Pharma KK

    Masahisa Jinushi

    Executive Director, Head of Medical Affairs, Gilead Sciences KK

    Shintaro Sengoku

    Professor, School of Environment and Society, Institute of Science Tokyo

    Hirokazu Shimoda

    Director, Bio-Industry Division, Commerce and Service Industry Policy Group, Ministry of Economy, Trade and Industry

    Yoshiki Sawa

    President,Organization of Future Medicine

    Toshio Fujimoto

    Chief Executive Officer, iPark Institute Co., Ltd.

    Yusaku Katada

    CEO, Restore Vision Inc.

    Morten Sogaard

    Head, Innovation Lab, Astellas Pharma

    Hiroo Igarashi

    President & Representative Director, Pfizer Japan Inc.

    Masamitsu Harata

    Chairman, CEO and Founder at Human Life CORD Japan Inc.

    Kotaro Yokote

    President, Japan Society for the Study of Obesity/ President, Chiba University

    Ataru Igarashi

    Associate Professor, Dept. of Public Health and Health Policy, Graduate School of Pharmaceutical Sciences, The University of Tokyo

    Gaku Hashimoto

    Professor, Faculty of Health and Welfare Services Administration, Kawasaki University of Medical Welfare (Former Member of the House of Representatives)

    Wataru Sugiura

    Director General, Center for Clinical Science, Japan Institute for Health Security

    Hiroto Araki

    Director, Infectious Diseases Control Division, Ministry of Health, Labor and Welfare

    Kennet Brysting

    President and Representative Director of Gilead Sciences KK

    Toshihiko Takeda

    Senior Advisor, Boston Consulting Group

    Kanae Kurata

    Director, Life Sciences Division, Research Promotion Bureau, Ministry of Education, Culture, Sports, Science and Technology (MEXT)

    Shin Kaneko

    Professor, Center for iPS Cell and Research Application, Kyoto University

    Kaneshiro Kiyohumi

    Chief Financial Officer, PeptiDream Inc.

    Takeyuki Akiyama

    Director, Japan Alliance for Lysosomal Disease Patient Organizations

    Shinichiro Komoto

    Eight Roads Ventures Japan・Partner

    Hiroshi Miyake

    Chief Executive Officer, Chordia Therapeutics

    Aya Kubota

    Editor in Chief, Nikkei Biotechnology & Business, Nikkei Business Publications, Inc.

    Jin Shiomura

    Founder, Managing Director & CEO of Nobelpharma

    Sumito Nishidate

    Chairman of the GIST & Sarcoma Patients and Families Association "NPO GISTERS"

    Kazuhiko Mori

    Senior Managing Director of Japan Pharmaceutical Manufacturers Association

    Masami Sakoi

    Chief Medical and Global Health Officer of Ministry of Health, Labor and Welfaresuka Miyabashira

    President of Japan Pharmaceutical Manufacturers Association

    Yasushi Okuno

    Professor, Department of Biomedical Data Intelligence, Graduate School of Medicine, Kyoto University

    Akira Izumi

    CEO & Founder, RyuWell Co.,LTD

    Hidenobu Ishizaki

    Executive director, Center for Development of Advanced Cancer Therapy, Cancer Institute Hospital, Japanese Foundation for Cancer Research (JFCR)

    Hiroyoshi Toyoshiba

    Director/ Chief Technology Officer

    Hiroyuki Tsunoda

    Deputy Head of Research Division, Chugai Pharmaceutical Co., Ltd.

    Masanori Shindo

    Deputy General Manager, Life Intelligence Consortium general incorporated association

    Sarah Brennan

    Company Group Chairman, Global Commercial Strategy Organization, for Johnson & Johnson Innovative Medicine

    Takahiko Iwaya

    Chair of European Federation of Pharmaceutical Industries and Associations, Japan (EFPIA Japan)

    Hans Klemm

    Japan Representative, Pharmaceutical Research and Manufacturers of America (PhRMA)

    Hitoshi Nakagama

    Japan Agency for Medical Research and Development/President

    Kazuto Yamada

    President and CEO, Japan Tissue Engineering Co., Ltd.

    Kiyoshi Ando

    Senior Staff Writer, Nikkei Inc.

  • up

    chris macrae

    Chartering is Q&A format we first publiished nearly 30 years ago - we were designing a process needed by organsiations aiming to serve f y64 352e a purpose that would be uniquely missed if they did not exist

    Today, it feels as if intelligence could make the whole world better or worse. Which handful of intelligence designers do you most need to be influenced by. Please note in chartering we aim to avoid listing 2 people with very similar impacts. If you think we have missed someone out with an essential impact please tell us chris.macrae@yahoo.co.uk

    AI Corporate Wizards 1 Jensen Huang -over 30 years now jensen and coworkers have redesigned computing to be a million + times more efficient for deep learning (and multidimensional) analysis which 21st C breakthroughs in AI depend on. WE can catalogue50 deepest partners of nvidia in different ways- countries: data sovereigntly; digital platforms eg self-drving cars in one of about 10 platforms that most majors connect through nvidia; supplying differentiated models as eg us 6 major digital comp[anoies (meta apple microsoft amazon google ,,,) become digital ai companies; full stack partnership eg from dell supplying devices at the edge to various digital twin services; links to ecosystems supporting 2000 startups in 100 countries; deep data partnerships eg with snowflake for clients of ai data warehousibg to service plus to safes force

    There are 2 particula engineering circles which huang appears to be in middle of. 1 building 25X more energy efficient supercomputers. Which will be first 10 countries with these and what will each supercomputer specialise in. Ai foundriies - ed desiging ai for clients. Most of the world’s best engineers for doing this are in Taiwan. Whilst digital twinning will likely show what skillsare needed if a franchise is to move its supply cains to another hemisphere o0f the world, 50 years of advanced manufacturing skills (eg demin quaslity), accelerating value leaps every few months) involes webs of hundered of just in time suppliers. Its not obvious how many engineers beyomd jensen have all the truts relationships needed to do this.

    AI Corperate Wizard 2 Demis was effectively the first to map pattern ai relevant to most of natural science challenges Einstein first -published E=mcsquared in 1905.With neumann einstein turing dyting suddenly in mod 1950s its almsot as if most of their wos forgotten untl hassabis rediscovered this early in 21st c. Back in 2012 google acquired hassabis deep mind when it was clear his training of mnachine learning ihn gmes lile go was going to achieve leaps like open sourcing 250 million plroteins. Thhis work continuens mainly out of cambridge and london offiecs whilst demi is als key to google own leaps into ai. One partnership google deep mind, nvidia, and disney todays owners of pixar is newton robotics. Codibg pixels with pixar when owned by steve jobs circa 2002 was arguably where nvidia first saw ai as the nain ourpose of the company. Today nvidia not ponly codes pixels but increasingly beleives ai will soon blurv line between programmking and simply defining what problem you want help withy in normal human convrerssation

    Yann Lecun is the most connected academic in ai and what were once the famous 3 of algorith wplrd (ie Hinton and bengio). Hinton has retired; bengio and lecun do summer camps. But lecun who first proved ai can do zipcodes in early 90s connects his chair pout of nyu courant maths lab wityh meta models from west coast to east coast tro france and india - egde o[pen ai models like llama and mistral are While the sacling iof llama by meta and mistral by mistral - the valiadtion as open academic models came from Lecun. The timing was good as the ai world series summit started by king charles has through France Macron and Modi india ended the drag onAI caused by EU bureaucracy.

    Country sponsor -if you speak english then King Charles support of AI as the most important innovation was timely; french modi; india - we are talking about ai agenting youth brains and productivie life not just big corporate ai nor gov ai by a few superpowers

  • up

    chris macrae

    Deep%20Compute%20Resources%20Comparison.docx

    Axios AI+

    By Ina Fried · Jun 24, 2025

    Apologies for missing that yesterday was National Typewriter Day, an embarrassing oversight for someone who has three actual and two Lego typewriters. Today's AI+ is 1,090 words, a 4-minute read.

     
     
    1 big thing: Musk's thumb on history's scale
    By 
     
    Illustration of a thumb pressing down on an arrow cursor.

    Illustration: Aïda Amer/Axios

     

    Elon Musk still isn't happy with how his AI platform answers divisive questions, pledging in recent days to retrain Grok so it will answer in ways more to his liking.

    Why it matters: Efforts to steer AI in particular directions could exacerbate the danger of a technology already known for its convincing but inaccurate hallucinations.

    The big picture: Expect more of Musk's thumb-on-the-scale approach, as governments and businesses build and embrace AI models with preferred responses on hot-button topics from LGBTQ rights to territorial disputes.

    Driving the news: In a series of tweets over the past week, Musk has expressed frustration at the ways Grok was answering questions and suggested an extensive effort to manipulate its output.

    • "We will use Grok 3.5 (maybe we should call it 4), which has advanced reasoning, to rewrite the entire corpus of human knowledge, adding missing information and deleting errors," he wrote on Saturday. "Then retrain on that. Far too much garbage in any foundation model trained on uncorrected data."
    • Musk also put out a call for people to suggest things that are "divisive facts," adding that he meant things that are "politically incorrect, but nonetheless factually true." The suggestions, though, included examples of Holocaust denialism and other conspiracy theories.
    • An xAI representative did not respond to a request for comment.

    Reality check: AI models are already hallucinating in ways that suggest failed attempts by company staff to manipulate outputs.

    • Last month, Grok started injecting references to "white genocide" in South Africa to unrelated conversations, which the company later attributed to an "unauthorized change" to its system.
    • At the other end of the political spectrum, Google and Meta appeared to make an effort to correct for a lack of diversity in image training data, which resulted in AI generated images of Black founding fathers and racially diverse Nazis.

    Between the lines: These early stumbles highlight the challenges of tweaking large language models, but researchers say there are more sophisticated ways to inject preferences that could be both more pervasive and harder to detect.

    • The most obvious way is to change the data that models are trained on, focusing on data sources that align with one's goals.
    • "That would be fairly expensive but I wouldn't put them past them to try," says AI researcher and Humane Intelligence CEO Rumman Chowdhury, who worked at Twitter until Musk dismissed her in November 2022.
    • AI makers can also adjust models after training them, using human feedback to reward answers that reflect the desired output.
    • A third way is through distillation, a popular process for creating smaller models based on larger ones. Creators can take the knowledge of one model and create a smaller one that aims to offer an ideological twist on the larger one.

    What they're saying: AI ethicists say the problem extends well beyond Musk and Grok. Many companies have been exploring how they can tweak answers to appeal to users, regulators and other constituencies.

    • "These conversations are already happening," Chowdhury tells Axios. "Elon is just dumb enough to say the quiet part out loud."
    • Chowdhury says his comments should be a wakeup call that AI models are in the hands of a few companies with their own set of incentives that may differ from those of the people using their services.
    • "There's no neutral economic structure," Chowdhury says, suggesting that instead of asking companies to "do good" or "be good," perhaps powerful AI models should be treated similar to utilities.

    Yes, but: Efforts to scour all bias from generative AI are doomed because the human data AI trains on is full of bias.

    • The training data reflects biases based on whose perspectives are over or underrepresented. There's also a host of decisions large and small made by model creators as well as other variables.
    • Meta, for example, recently said it wants to remove bias from its large language models, but experts say the move is more about catering to conservatives than achieving some breakthrough in model neutrality.

    Bottom line: Ultimately — as we reported over a year ago — it boils down to a battle over what values powerful AI systems will hold.