stephtara@indieweb.social (@stephtara) profile picture

Anglo-Swiss. My blog is still alive. @stephtara@indieweb.social when this ship is done sinking. ADHD, cats, judo, sailing, people, communities. Unfacebooked...

Lausanne, Switzerland
Joined December 2006
Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)Photo by stephtara@indieweb.social (@stephtara)
I've been saying this for something like eight months now, with varying approaches and volume levels, and no one is paying attention yet. This is either because I am way early or my idea isn't one third as good as I think it is. :) But here it goes again: I think we will soon figure out that the entire game for AI is articulation of ideal state. And that all of our conversations about specs, and plans, and loops, and PRDs and so many other things that have been talked about in software engineering and harness engineering for the last six months...all converge on this same idea. I think the way it will be articulated is in the form of a single artifact that captures, enhances, iterates on, climbs toward, builds, and tests the ideal state. One document that replaces all these loops and specs and PRDs and plan files and design files, et cetera. My current implementation of this is the ideal state articulation (ISA) system. It turns prompt engineering into intent engineering, in the sense that it abandons telling the AI how to do things and replaces that with telling it exactly what you want the output to be. @karpathy had a recent post where you talked about going on a walk and just talking freely through a bunch of ideas related to what you're trying to accomplish. to me, that is 100% on point and 100% part of this ideal state articulation. We also had this recent open AI hack where a goal was given and a goal was presumably achieved. Similar to a paperclip maximizer. And in both cases, the problem is that we conveyed a goal without conveying ideal state. Which, if properly articulated, would have included lots of paperclips but humanity still existing or getting the top grade on the test, but without building new exploits and hacking companies. The problem is always in the guessing. It's in the gap between stated and implicit. Or stated and difficult to articulate. That's why I see the whole entire game as this articulation of what we want. This is what prompt engineering always has been. But it got muddled up with earlier models that weren't very smart, so we had to combine the intent engineering with step-by-step instructions. I think it's time to cut through that now and isolate this central idea. An extraordinary amount of our problems, really across anything but especially when working with others and building things with AI, is assuming that the receiver has the same idea in their mind as the one in ours. And that's why I'm proposing this approach for addressing that issue directly. Let's get extremely good at articulating the ideal state for what we have asked for. Really nailing this does so many things for us simultaneously. It not only helps the receiving system build what we actually want with a lot less gap between our mind and its mind, but after it's built, it also becomes the testing harness as well as the heart of the documentation. It's a single system for general hill-climbing, toward anything. docs.ourlifeos.ai/Isa__IsaSy…
25
22
6
133
14,937
Ok this is just registering for me that this is the government demanding that Anthropic not allow Foreign Nationals to use Fable. So they’re not saying the model is generally too dangerous; they’re saying it’s too good for Foreign Nationals to have access. And since Anthropic can’t instantly do that, even if they wanted to, they had to disable access for everyone for now.
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…
86
28
4
353
83,604
Meta illegaly downloaded 80+ terabytes of books from LibGen, Anna's Archive, and Z-library to train their AI models. Aaron Swartz downloaded 70 GBs of articles from JSTOR (0.0875% of Meta) in 2010. Faced $1 million in fine and 35 years in jail. Took his own life in 2013.
101
3,335
102
11,700
409,580
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
The degree to which you are awed by AI is perfectly correlated with how much you use AI to code.
1,262
2,703
718
22,760
7,651,369
🇵🇸 « Nous journalistes à Gaza, sommes des condamnés à mort » explique le journaliste Rami Abu Jamous à Alice Froussard. Depuis le 7 octobre 2023, l'armée israélienne a tué au moins 242 journalistes palestiniens selon le Haut-Commissariat des Nations unies aux droits de l'Homme.
9
153
3
189
15,339
Some operatives from the Department of Government Efficiency (DOGE) are joining Airbnb cofounder and DOGE member Joe Gebbia’s new National Design Studio, sources tell WIRED. "My directive is to update today’s government services to be as satisfying to use as the Apple Store" wired.com/story/doge-operati…
3
11
2
43
33,641
NEW: @WhatsApp caught & fixed a sophisticated zero click attack... Now they've published an advisory about it. Say attackers combined the exploit with an @Apple vulnerability to hack a specific group of targets (i.e. this wasn't pointed at everybody) Quick thoughts 1/
24
199
14
604
200,475
You might assume that sex ed is something that has gotten progressively better over time, but you'd be wrong. We've been stuck in the same pattern and cycle of debate about sex ed for over 100 years--and today, it's as controversial as ever. sexandpsychology.com/blog/po…
2
5
697
Was at the #WCUS talk today on the Core AI roadmap for WordPress. I like how focused on experimentation in the name of eventual standardization (and moving quickly) they are being If you haven't seen their approach detailed yet, read this first: make.wordpress.org/ai/2025/0… Breaking it down 👇
1
7
16
2,744
2/2 But it also shows the bias that professional women always have to deal with. I have huge respect for Dr. Rasmussen and what she does for public education on vaccines and public health. If NPR can't get it right...
1
5
47
942
Today on 1A @NPR featured an interview with Dr. @angie_rasmussen and @DrPaulOffit . Both were excellent as always. But the host always referred to Dr. Rasmussen as "Angie" and Dr. Offit as "Dr. Offit", never using her formal title. Her title is important, she is an expert. 1/2
9
19
1
113
8,522
Le plus dur ça sera pour les gens qui vont atteindre l'âge de la retraite d'ici 10 ans. Eux auront payé plein pot toute leur vie pour les boomers, mais leurs retraites seront minimales. Bien plus dur que pour les boomers, même si on reformat aujourd'hui. – at Levallois-Perret, France
214
96
24
1,374
93,190
BREAKING: At least 3 CDC leaders have resigned tonight: Dan Jernigan, National Center for Emerging and Zoonotic Infectious Diseases Deb Houry, chief medical officer Demetre Daskalakis, National Center for Immunization and Respiratory Diseases Possibly more TK
NEW: Following the reports of Monarez’s departure as @CDCgov director, several other leaders have just submitted their resignations: Dan Hernigan, head of the National Emerging and Zoonotic Infectious Diseases and Deb Houry, the CDC’s chief medical officer. Unclear if related.
759
2,530
466
6,004
3,175,287
Excellente nouvelle : la SNCF colle une amende de 110 euros à la propriétaire d'un chat qui miaule dans un train On va enfin pouvoir demander de telles verbalisations pour les enfants qui crient ou les bébé qui pleurent La SNCF toujours aussi merdique ow.ly/6uYY50WMsCE
183
872
35
2,896
63,442
My resignation letter from CDC. Dear Dr. Houry, I am writing to formally resign from my position as Director of the National Center for Immunization and Respiratory Diseases at the Centers for Disease Control and Prevention (CDC), effective August 28, 2025, close of business.   I am happy to stay on for two weeks to provide transition, if requested. This decision has not come easily, as I deeply value the work that the CDC does in safeguarding public health and am proud of my contributions to that critical mission. However, after much contemplation and reflection on recent developments and perspectives brought to light by Secretary Robert F. Kennedy Jr., I find that the views he and his staff have shared challenge my ability to continue in my current role at the agency and in the service of the health of the American people. Enough is enough. While I hold immense respect for the institution and my colleagues, I believe that it is imperative to align my professional responsibilities to my system of ethics and my understanding of the science of infectious disease, immunology, and my promise to serve the American people.  This step is necessary to ensure that I can contribute effectively in a capacity that allows me to remain true to my principles. I am unable to serve in an environment that treats CDC as a tool to generate policies and materials that do not reflect scientific reality and are designed to hurt rather than to improve the public’s health.  The recent change in the adult and children’s immunization schedule threaten the lives of the youngest Americans and pregnant people.   The data analyses that supported this decision have never been shared with CDC despite my respectful requests to HHS and other leadership.  This lack of meaningful engagement was further compounded by a “frequently asked questions” document written to support the Secretary’s directive that was circulated by HHS without input from CDC subject matter experts and that cited studies that did not support the conclusions that were attributed to these authors.  Having worked in local and national public health for years, I have never experienced such radical non-transparency, nor have I seen such unskilled manipulation of data to achieve a political end rather than the good of the American people. It is untenable to serve in an organization that is not afforded the opportunity to discuss decisions of scientific and public health importance released under the moniker of CDC.  The lack of communication by HHS and other CDC political leadership that culminates in social media posts announcing major policy changes without prior notice demonstrate a disregard of normal communication channels and common sense.  Having to retrofit analyses and policy actions to match inadequately thought-out announcements in poorly scripted videos or page long X posts should not be how organizations responsible for the health of people should function.  Some examples include the announcement of the change in the COVID-19 recommendations for children and pregnant people, the firing of scientists from ACIP by X post and an op-ed rather than direct communication with these valuable experts, the announcement of new ACIP members by X before onboarding and vetting have completed, and the release of term of reference for an ACIP workgroup that ignored all feedback from career staff at CDC. The recent term of reference for the COVID vaccine work group created by this ACIP puts people of dubious intent and more dubious scientific rigor in charge of recommending vaccine policy to a director hamstrung and sidelined by an authoritarian leader.   Their desire to please a political base will result in death and disability of vulnerable children and adults.  Their base should be the people they serve not a political voting bloc. I have always been first to challenge scientific and public health dogma in my career and was excited by the opportunity to do so again.  I was optimistic that there would be an opportunity to brief the Secretary about key topics such as measles, avian influenza, and the highly coordinated approach to the respiratory virus season.  Such briefings would allow exchange of ideas and a shared path to support the vision of “Making America Healthy Again.”  We are seven months into the new administration, and no CDC subject matter expert from my Center has ever briefed the Secretary.  I am not sure who the Secretary is listening to, but it is quite certainly not to us.  Unvetted and conflicted outside organizations seem to be the sources HHS use over the gold standard science of CDC and other reputable sources.  At a hearing, Secretary Kennedy said that Americans should not take medical advice from him.  To the contrary, an appropriately briefed and inquisitive Secretary should be a source of health information for the people he serves. As it stands now, I must agree with him, that he should not be considered a source of accurate information. The intentional eroding of trust in low-risk vaccines favoring natural infection and unproven remedies will bring us to a pre-vaccine era where only the strong will survive and many if not all will suffer.  I believe in nutrition and exercise.  I believe in making our food supply healthier, and I also believe in using vaccines to prevent death and disability.  Eugenics plays prominently in the rhetoric being generated and is derivative of a legacy that good medicine and science should continue to shun. The recent shooting at CDC is not why I am resigning.  My grandfather, who I am named after, stood up to fascist forces in Greece and lost his life doing so.  I am resigning to make him and his legacy proud.   I am resigning because of the cowardice of a leader that cannot admit that HIS and his minions’ words over decades created an environment where violence like this can occur.  I reject his and his colleagues’ thoughts and prayers, and advise they direct those to people that they have not actively harmed. For decades, I have been a trusted voice for the LGBTQ community when it comes to critical health topics.  I must also cite the recklessness of the administration in their efforts to erase transgender populations, cease critical domestic and international HIV programming, and terminate key research to support equity as part of my decision. Public health is not merely about the health of the individual, but it is about the health of the community, the nation, the world. The nation’s health security is at risk and is in the hands of people focusing on ideological self-interest. I want to express my heartfelt gratitude for the opportunities for growth, learning, and collaboration that I have been afforded during my time at the CDC. It has been a privilege to work alongside such dedicated professionals who are committed to improving the health and well-being of communities across the nation even when under attack from within both physically and psychologically. Thank you once again for the support and guidance I have received from you and previous CDC leadership throughout my tenure. I wish the CDC continued success in its vital mission and that HHS reverse its dangerous course to dismantle public health as a practice and as an institution.  If they continue the current path, they risk our personal well-being and the security of the United States. Sincerely, Demetre C. Daskalakis MD MPH (he/his/him)
21,512
15,135
3,987
54,320
19,800,106
Salut! Mon prochain livre est en précommande pour le distribuer encore mieux!!!! Sortie prévue le 23 octobre! Le but c'est d'apprendre à ne plus subir les autres!!!! On en reparle bientôt! youtu.be/8EO6PZJkwTQ Lien de préco: fnac.com/a21813826/
1
9
31
8,677
it should probably be a bigger story that most of the world has stopped shipping packages to the U.S. entirely because of Trump’s tariffs
2,302
17,014
1,630
86,718
7,393,848
my mom just found a new use case for chatgpt
326
1,057
304
20,272
1,213,439