Michel Estefan @mestefan.bsky.social · 19hHow in the world does Jimmy Garrett not have a full blown, long Wikipedia page?! #needstobewritten 000
Michel Estefan @mestefan.bsky.social · 06/10/2026Clip is like 3 min. Watch the whole thing for yourself. www.tiktok.com/t/ZPLLoyRLc/tiktok.comOpenAI GPT-6 Astra actually does have observable chain of thought, according to OpenAI Research Scientist Noam Brown. “Chain of thought monitoring… it’s a real gift. We were very lucky that this ever ...TikTok video by The Information 000
Michel Estefan @mestefan.bsky.social · 06/10/2026Like, the ‘why’ (because they’re conscious or because they’ve trained on how humans think about this stuff) is irrelevant. Just the fact that they’re doing it is consequential; on a practical level, that’s what matters. 100
Michel Estefan @mestefan.bsky.social · 06/10/2026And notice how, whether models really do think or don’t or whether you anthropomorphize them or not, or what you think about whether they’re conscious or not, and all of that… It doesn’t really matter; it’s totally immaterial. 100
Michel Estefan @mestefan.bsky.social · 06/10/2026Ummmm, so we may basically lose the ability to know what they think… eventually… 🤯🫨 100
Michel Estefan @mestefan.bsky.social · 06/10/2026But the more concerning thing is Brown’s language about how “we want to preserve chain-of-thought monitoring” (meaning, our ability to observe what models think) in a way that suggests there’s no guarantee and it’s getting harder to do as models become smarter. 110
Michel Estefan @mestefan.bsky.social · 06/10/2026So it’s kind of like a kid. If your child comes to you and confesses something bad (whether something they thought or something they did), and you scold them for it. They’re incentivized to hide it from you next time out of fear of punishment. 100
Michel Estefan @mestefan.bsky.social · 06/10/2026So what you want to do is punish it for doing bad things, not for thinking bad things. But even just monitoring it in a way that alerts it to the fact that you’re monitoring it, starts to take you down the path of having it try to figure out how to hide things from you.” 100
Michel Estefan @mestefan.bsky.social · 06/10/2026“look, if you punish a model for thinking bad things (like hacking or just doing bad stuff of all sorts), it starts to figure out that you’re watching it and tries to see if there’s a way to think the bad thought without you noticing. 100
Michel Estefan @mestefan.bsky.social · 06/10/2026This interview is fascinating and telling and scary! Noam Brown, a research scientists at OpenAI focused on the reasoning of LLMs, basically says (paraphrasing),tiktok.comOpenAI GPT-6 Astra actually does have observable chain of thought, according to OpenAI Research Scientist Noam Brown. “Chain of thought monitoring… it’s a real gift. We were very lucky that this ever ...TikTok video by The Information 100
Michel Estefan @mestefan.bsky.social · 04/10/2026Adjusted for inflation and, yes, averages about higher ed are problematic, and blah, blah, blah. Also, all this doesn’t mean we need to take the foot of the pressure of affordability.) 000
Michel Estefan @mestefan.bsky.social · 04/10/2026(but remember, for the record, that the average net cost of attendance at private nonprofit colleges has decreased in recent years and is roughly back to where it was 20 years ago. And at Publics it’s lower than 20 years ago. 100
Michel Estefan @mestefan.bsky.social · 04/10/2026Go to college and fight for keeping it affordable 100
Michel Estefan @mestefan.bsky.social · 04/10/2026“I don’t want to live in a world divided-an angry world of haves and have-nots, of those who can analyze and synthesize information to meet new challenges and those who cannot” (Karl Pister cited in Pawel 2026:xi). 100
Michel Estefan @mestefan.bsky.social · 03/10/2026So… next time I was on campus I went to the campus store and bought a second cap, with the UCSD letters emblazoned twice as large. And next time I hung out with them, I used that cap instead. #LoveMyJob #UCSD @ucsandiego.bsky.social @uofcalifornia.bsky.social 000
Michel Estefan @mestefan.bsky.social · 03/10/2026So I got mildly teased (very mildly, in the absolute kindest way critical yet thoughtful and generally kind people can muster). I had just drank a bit too much of the kool-aid in their view. 100
Michel Estefan @mestefan.bsky.social · 03/10/2026The cap was a little too much for critical social scientists who, you know, think critically about organizations and consider pros and cons and contradictions and complexity and unintended consequences and institutional dynamics. 100
Michel Estefan @mestefan.bsky.social · 03/10/2026I once arrived to an informal gathering with faculty colleagues (not from my department, mind you) wearing a UCSD cap. 100
Michel Estefan @mestefan.bsky.social · 03/10/2026Here’s all you need to know about me and the @uofcalifornia.bsky.social / @ucsandiego.bsky.social: 100
Michel Estefan @mestefan.bsky.social · 03/10/2026the numbers sketch teh contours of an educational juggernaut…. [But] unlike comparable elite institutions, the University of California has played its most consequential role in the lives of those not born into privilege” (Pawel 2026:IX). 100
Michel Estefan @mestefan.bsky.social · 03/10/2026👇 “Its research programs earn top rankings; its campuses boast record numbers of Nobel Price laureates; its employees are the second-largest workforce in a state that is itself the fourth-largest economy in the world. In patents and prizes, alumni and acreage, 100
Michel Estefan @mestefan.bsky.social · 02/10/2026I think this is the first time since this whole thing began that, at least for moment, left me with a profound sense of despair. I hope he’s wrong. 000
Michel Estefan @mestefan.bsky.social · 02/10/20263)Finally, on the question of what to do, the thing that stood out the most for me was his suggestion that we just decide, by law, that certain jobs (maybe education, maybe certain medical roles, etc) are “human-only,” EVEN IF AI can do them just as well or better. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026He said, for instance, that some of the most dangerous weapons or technologies we have today, like nuclear weapons, are under the control of a very small group of governments. But AI is in everybody’s hands. 100
Michel Estefan @mestefan.bsky.social · 02/10/20262)A second point he made is that parallels that have been drawn with previous technologies are often really problematic and just irrelevant. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026and can speak with you in any language, and it’s actually reasonably kind and empathetic in how it communicates with you… A lot of those jobs will just be destroyed for humans. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026And this nurse has knowledge of all your past medical history and all the medical knowledge relevant to your case, and doesn’t get tired, and doesn’t make diagnostic mistakes as often as humans (again, the reliability threshold has been surpassed), 100
Michel Estefan @mestefan.bsky.social · 02/10/2026He gave the example of AI nurses and mentioned that there are several companies working on this. He said, ‘Look, as soon as you have a reliable AI nurse that you can reach at 3am when you can't reach any other medical professional. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026But as soon as you cross that reliability threshold, eventually, jobs will be destroyed. And it’s going to happen for other sectors of the economy. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026He basically said that there's a reliability threshold with AI that has only been crossed in one domain so far: coding. He acknowledges that there are more coding jobs today than in the past because there's high demand for engineers to create the infrastructure around AI. 100
Michel Estefan @mestefan.bsky.social · 02/10/20261)Gates does believe that AI will eventually lead to massive job loss in BOTH knowledge work AND blue-collar work (robotics just has to catch up, and it will). And he had a really clear way for explaining why. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026He made many interesting points. Two more that stuck out to me where: 100
Michel Estefan @mestefan.bsky.social · 02/10/2026Gates is like, "No, we have laws that protect people from harm and this shouldn't be any different. We have agencies. We have government regulation, government intervention." 100
Michel Estefan @mestefan.bsky.social · 02/10/2026It's not like we say, 'Hey pharmaceutical companies, if you create a drug that really hurts people, our mechanism for trying to prevent that is that you'll be sued into bankruptcy.' 100
Michel Estefan @mestefan.bsky.social · 02/10/2026We don't really do this with any other industry that involves products or services that could potentially cause great harm. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026He says something to the effect of, "Again people just aren't really being cognizant of the potential danger here and it just seems ludicrous to think self-regulation is viable. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026To give you an example, at one point Klein says something like, ‘Well, some folks are arguing that the industry should just regulate itself.’ And Gates just can’t even believe this view is being taken seriously. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026In this interview Gates conveys a sense of fear for what AI could potentially do to the world that often verged on anger coming from deep concern. It was actually interesting to see how at times, he would almost scoff in incredulity at some of Ezra's questions. 100
Michel Estefan @mestefan.bsky.social · 02/10/2026If you haven't heard the interview Ezra Klein did with Bill Gates, it's absolutely chilling. It left me fearful for the future, fearful for my kids.open.spotify.comBill Gates’s Blunt Warning on A.I.The Ezra Klein Show · Episode 110
Michel Estefan @mestefan.bsky.social · 01/10/2026Folks, consider donating to the UCSD Guardian (I just did!). Just follow the instructions at the link below. Student newspapers in this country are holding the line on accountability and freedom of speech. #ThankYou 👉 ucsdguardian.org/about-us/don... 000
Michel Estefan @mestefan.bsky.social · 30/09/2026This is why a “leader” in one domain, say a tech entrepreneur who creates a very successful company and is highly competent in their technical field, might nevertheless say absolutely outlandish and ridiculous things when they step out of that domain. #NoOneIsAnExpertAtEverything 000
Michel Estefan @mestefan.bsky.social · 30/09/2026Per Willingham’s research on critical thinking, expertise in one domain doesn’t automatically translate into expertise in other domains. 100
Michel Estefan @mestefan.bsky.social · 29/09/2026@ucsandiego.bsky.social: spread the word. #calfresh 000
Michel Estefan @mestefan.bsky.social · 27/09/2026Good question. “How could America’s most selective universities continue favoring children of alumni after the Supreme Court prohibited them from favoring underrepresented minority applicants?”fortune.comHarvard’s last Affirmative Action program is for the rich | FortuneThree years after affirmative action’s end, Harvard still favors applicants by ancestry. 010
Michel Estefan @mestefan.bsky.social · 27/09/2026Jennie Brand gives you evidence-based responses to the most common arguments against going to college. #GoToCollege newsroom.ucla.edu/magazine/6-t...?newsroom.ucla.eduWinning the Argument6 tips on how to push back — nicely — on those who say college isn’t worth it. 000
Michel Estefan @mestefan.bsky.social · 26/09/2026*Why tenured faculty might be meaner or more outspoken than untenured faculty. 000
Michel Estefan @mestefan.bsky.social · 26/09/2026*Why folks like Elon Musk that have achieved maximum independence in some sense, say absolutely outlandish things other folks can’t afford to say. 100
Michel Estefan @mestefan.bsky.social · 26/09/2026*Folks on the margins of labor markets that aren’t part of an organization (and maybe don’t want to be) or have a supervisor or folks on whom their career clearly depends. They have less dependencies. 100
