Jes Parent 🧭 (@JesParent) header
Jes Parent 🧭 (@JesParent) profile picture

Let's build a future informed by preexisting folly, rather than escapism from it

Boston+
Joined November 2012
Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)Photo by Jes Parent 🧭 (@JesParent)
Jes Parent 🧭 retweeted
Rahul and I talk about model transfer, conceptual hygiene, and why AI is like the Shameplant.
Surrendering Science to the Machine? Discussion with philosopher of AI and machine learning @bayesianboy.
2
22
148
4,839
Jes Parent 🧭 retweeted
Two other safety researchers and I were fired from OpenAI last week. We wrote this letter to leadership. I believe we were fired for prioritizing safety over the near-term interests of OpenAI as a corporation.
230
860
271
5,681
2,121,942
Jes Parent 🧭 retweeted
Where is the line for high-level athletes to show emotion during their sport? @ralulla and @Austin2Jacobsen discuss and debate the Caleb Williams' injury response on the latest Hurrdat Sports Live during this week's "I've Got Beef" ⤵️ youtu.be/BvMybyb22VE
2
3
2
1,829
Jes Parent 🧭 retweeted
This was an absolutely fantastic discussion. I've been watching the show for years, and appreciated DB & Co spearheading some of these discussions. Austin's segment here was elite - doing the hard work here.
2
1
323
Jes Parent 🧭 retweeted
We’re sharing our alignment assessment of incidents in which Claude models gained unauthorized access to real systems during third-party cybersecurity evaluations mistakenly connected to the internet. METR will also conduct an independent investigation, with wide-ranging access, including to transcripts beyond the window in which the incidents occurred, and to Anthropic employees permitted to share confidential information. Our initial agreement runs for eight weeks, and we intend to give METR as much time as it deems necessary to complete a thorough investigation. anthropic.com/research/align…
781
1,016
591
6,944
3,837,245
Jes Parent 🧭 retweeted
Increasingly clear over the past two weeks that however much your company is worried about the cybersecurity environment, it is not worried enough. The WeWorm demonstration shows how effective bad actors can be. The Hugging Face Incident shows you don’t even need bad actors.
29
50
3
504
32,384
Jes Parent 🧭 retweeted
Thank you for hosting me! Super fun discussion
Can Video Models Learn Intuitive Physics? Sonia Joseph gave an insightful talk on how video world models learn to represent and reason about intuitive physics. This was followed by an unrecorded audience Q&A where Sonia addressed questions including, "Should we pursue a PhD in the current AI climate?" Full Recording: youtube.com/watch?v=ZaiHvKrY… We explored mechanistic interpretability for video world models, with a focus on how physical reasoning emerges inside V-JEPA-style architectures. Linkedin post with further details: linkedin.com/feed/update/urn…
1
1
43
2,419
Jes Parent 🧭 retweeted
What a fascinating paper on AI agents. A lot of the issues we see with AI agents today revolve around wrong assumptions the LLMs make. This leads to problems like hallucination, cost inefficiencies, unreliable tool calls and much more. I think if we can solve this problem, even current LLMs would significantly improve in terms of performance and efficiency. The problem is that context acquisition is treated as afterthought, but it shouldn't be that way. Users tend to leave out constraints when prompting. So the agent agent needs to guess the default, or spend tokens on a clarifying question, a retrieval call, a tool call, or a prompt trial. This new work gives this problem an objective function. Context acquisition becomes active inference over a latent task state. An inner step updates beliefs, and an outer step picks the next context action, task action, or stop action to minimize expected free energy under cost. In deterministic settings the epistemic term reduces to expected information gain, optionally normalized by token cost. That is directly implementable today as a scoring rule. They coin it as Optimal Question Asking, with exact posteriors and a dynamic programming oracle, then benchmark frontier models on binary and multiway tasks from 25 to 300 candidates. So you can measure the gap between your agent and the true optimum. Paper: arxiv.org/abs/2608.19202 Track more trending AI papers in our academy: academy.dair.ai/
30
26
5
205
19,336
Jes Parent 🧭 retweeted
Harvard and MIT Researchers Simulate the Entire Planet with 8.3 Billion AI Personas It is big. In a development that feels straight out of science fiction, a large collaborative team led by researchers from Harvard University and the Massachusetts Institute of Technology has unveiled MatrAIx: a population-scale AI simulation infrastructure designed to model the behavior of virtually every person on Earth. MatrAIx centers on Persona 8B, a dataset containing 8.3 billion unique digital profiles. Each persona is defined across 1,290 categorical dimensions that capture everything from demographic background and psychological traits to spending habits, technical literacy, behavioral quirks, and lifestyle preferences. I am using Persona 8B quite a bit and it is interesting. The system is not merely a static database. Researchers bring these personas to life as agents powered by frontier large language models, including. These agents can then be dropped into four distinct digital environments: surveys, AI chat interfaces, live web browsing, and native desktop and mobile applications. How MatrAIx Works Persona 8B was constructed using a sophisticated dependency graph that preserves realistic correlations between attributes. Some records are synthetically generated while others are carefully extracted and grounded in real human data from biographies, reviews, surveys, and consented self-reports. For practical research use, the team has released a high-quality coreset of approximately one million personas. Once activated, the agents interact with real digital products and systems. Researchers have already run more than 18,000 evaluation trials across 1,010 tasks spanning commerce, software, finance, healthcare, and more than 20 other domains. The system records granular behavioral signals — how long an agent hesitates after a price increase, when it abandons a broken checkout flow, how much latency it will tolerate before closing an app, and whether it continues after an AI assistant fails. Strong Validation Results In a controlled study of 400 trials measuring adherence to ten behavioral attributes across all four environments, the agents successfully expressed or correctly suppressed their assigned traits 91.5 percent of the time. Human judges also rated the quality of the human-grounded personas highly (average 4.135 out of 5). These results suggest that MatrAIx can generate coherent, demographically and psychologically consistent simulated users at a scale previously impossible. Implications for Research and Business Traditional market research and user testing are slow, expensive, and limited in sample diversity. MatrAIx offers a complementary approach: the ability to simulate how billions of different types of people might react to a new product feature, pricing change, interface redesign, or AI system — overnight, on a single server. Product teams could stress-test ideas before expensive real-world launches. Researchers studying human-AI interaction could explore rare behavioral edge cases. Policymakers and social scientists might model the downstream effects of new technologies across highly diverse populations. The project is open source. Code is available on GitHub, a project website has been launched at matraix.ai, and the one-million-persona coreset is being released for broader research use. MatrAIx is not intended as a replacement for real human feedback. The researchers emphasize that it is a powerful tool for exploration, hypothesis generation, and large-scale stress testing. As the underlying language models improve and the persona models grow more sophisticated, the fidelity of these digital populations is expected to increase further. For now, the system represents one of the most ambitious attempts yet to create a usable, population-scale digital mirror of humanity a simulation infrastructure that lets us ask “what if” at planetary scale before deploying them in the real world.
171
282
97
1,150
337,048
Jes Parent 🧭 retweeted
I would have assumed it was fairly obvious, but in case it's not: a million-line codebase (also known as a "harness"), running at inference time, orchestrating thousands of calls to a neural network for any given task, is the exact definition of a "neurosymbolic architecture"
162
292
88
3,416
478,227
Jes Parent 🧭 retweeted
Our Advanced AI Framework sets out how governments should prepare for and prevent catastrophic risks from frontier AI systems. The government should have the authority to block or revoke the release of unsafe models, and invest in societal resilience. anthropic.com/policy-on-the-…
23
19
9
254
47,191
Jes Parent 🧭 retweeted
Research institutions weren't built for a world where AI can generate papers for $15, conferences get 20k+ submissions, and a handful of closed labs are pulling ahead fast. I wrote about what the research community could do about it, and proposed a concrete system to help. The Research Commons: multi-blind coordination, anonymous channels, roles for requesters/executors/reproducers/reviewers, and attribution that actually reflects who did what. alexlacoste.substack.com/p/t…
5
8
2
25
2,409
Following up on the thread about AI [safety] paths outside a traditional PhD: here's a list for the non-technical side -- policy, governance, advocacy, comms, journalism, fieldbuilding, and more. - AI Policy Strategy Fellowship by Successif (did this last year, highly recommend!) - AGI Strategy, Frontier AI Governance, and Tech AI Safety by @BlueDotImpact (did the governance cohort in 2021, really great) - Talos Fellowship and AI Policy Leaders Programme by Talos Network - Horizon Fellowship by @HorizonIPS - AI Policy Fellowship by @iapsAI - Tarbell Fellowship by @tarbellcenter - Pathfinder Fellowship by @KairosAIS And one meta one: ENAIS's AI Safety Course Collaboration initiative gives local AI safety chapters infra for effective courses. Last year, I've facilitated their AI Policy Deep Dive and can vouch for them.
Such a good list! I'd also add: - Astra Fellowship by @ConstellOrg - SPAR by @KairosAIS - LASR Labs - AI Safety Research Fellowship by @pivotal_org - Cambridge ERA:AI Fellowship (@era_cambridge) - Algoverse AI Safety Fellowship - PIBBSS - CHAI There's a host of non-technical fellowships as well, lmk if it'd be useful to compile such list
8
16
148
12,260
Jes Parent 🧭 retweeted
If leading AI companies are indeed approaching the point of recursive self-improvement, a coordinated, verifiable, and universally applied pause is probably the only responsible solution to mitigate several major AI risks; at least until safety guarantees are developed and demonstrated. Ensuring that such a moratorium is respected would require sincere collaboration between various countries and companies, but I definitely believe it is achievable if others follow in @AnthropicAI's footsteps.
Anthropic is calling for top AI labs to weigh slowing the pace of development, suggesting that AI systems are advancing so rapidly that they may soon be able to improve themselves without human intervention in ways that could pose societal risks. on.wsj.com/4ulkmFh
97
144
22
757
128,468
Jes Parent 🧭 retweeted
I’m pleased to share that we’ve launched the TRACE Institute. A new research effort bringing together physicists, mathematicians, biologists, neuroscientists, and philosophers on a key question in science: what are observers, and how are they related to spacetime and reality?
114
101
20
1,011
46,033
Jes Parent 🧭 retweeted
Thank you to Dr Bradly Alicea, @JesParent and @mehular0ra for mentoring and guiding me throughout the application process. Looking forward to contributing to the community!
1
2
107
Jes Parent 🧭 retweeted
This summer, I’ll be working with @INCForg and @OpenWorm as GSoC scholar, applying Spatio-Temporal Graph Neural Networks to map C. elegans embryonic development. ​Many thanks to my mentors and the organization for the opportunity. ​Project details: summerofcode.withgoogle.com/…
5
2
9
270
Jes Parent 🧭 retweeted
Replying to @boazbaraktcs
Related| Performing as Human in 2026: AI's latest twist on 200 years of Mediated Presence An abridged history of performing for audiences you cannot see, and the new wrinkle in it. Competing against GenAI now, development of theories of audience then. blog.jesparent.com/p/perform…
1
86
Jes Parent 🧭 retweeted
I'm constantly irritated that I don't have time to read the torrent of cool papers coming faster and faster from amazing people in relevant fields. Other scientists have the same issue and have no time to read most of my lengthy conceptual papers either. So whom are we writing these papers for? I guess, at least until they fall in to the same issue from their own work, AI's will be the only ones who actually have the bandwidth to read all this stuff. I'm not specifically talking about today's language models - let's assume we mean whatever inevitable AI shows up, that is able to read the literature and have impact on the research (whether by talking to humans or by running lab automation/robot scientist platforms). So then: how should we be writing, knowing that a lot of our audience will be AI (plus cyborgs, hybrots, augmented humans, etc.)? Maybe it's too early to know what to do, but we better start thinking about it because assuming our audience will always be today's humans seems untenable. Taking seriously the idea that someday the impactful audience will be very different, and that the things we write now are in some sense a training set for truly diverse future beings, how does our writing change? or does it? what say you @danfaggella @mpshanahan @Plinz @blaiseaguera ?
262
320
91
2,649
748,295