--- generator: contextjamming-llm-context title: "Sabrina Gonzalez Pasterski, PhD" description: "The kid who built a Zenith CH 601 XL became a co-architect of celestial holography — and now runs Project Theoria, an attempt to get AI and theoretical physics working on the same problem." url: https://www.contextjamming.com/founder-files/pasterski markdown_url: https://www.contextjamming.com/founder-files/pasterski.md section: "Founder Files" content_type: "FounderFile" author: "Bret Kerr" publisher: "Context Jamming (ACRA Insight LLC)" date_published: 2026-11-12 language: en-US word_count: 4511 image: https://pub-bcdb3a688294470c8056f904d7e6236b.r2.dev/FounderFile/sabrina-pasteeraski.png --- > **Sabrina Gonzalez Pasterski, PhD** — Founder Files, Context Jamming. This is a plain-Markdown copy of https://www.contextjamming.com/founder-files/pasterski, prepared for LLM context windows, agents and retrieval. > > Cite as: Bret Kerr, "Sabrina Gonzalez Pasterski, PhD," *Context Jamming*, November 12, 2026, https://www.contextjamming.com/founder-files/pasterski Asymptotic Symmetries and Detectors — Part 1 Institute for Advanced Study · Prospects in Theoretical Physics 2026 · TL;DR ↓ FounderFiles·N°074·Physics · Holography · AI for Science 2005 — ![Sabrina Gonzalez Pasterski — theoretical physicist, founder of Perimeter's Celestial Holography Initiative and Project Theoria (illustration)](https://pub-bcdb3a688294470c8056f904d7e6236b.r2.dev/FounderFile/sabrina-pasteeraski.png) *Fig. · The celestial cartographerIllustration* Subject·Sabrina Gonzalez Pasterski, PhD·Theoretical physicist · Founder, Celestial Holography Initiative (Perimeter) · Founder, Project Theoria # Sabrina Gonzalez Pasterski, PhD She built an airplane before she could drive, helped move the physics of the infrared from a curiosity into a research program, and is now trying to get AI to read the whole of theoretical physics. *At 14 she soloed a Cessna 150 in Sarnia, Ontario. At 21 she co-discovered a new kind of gravitational memory. At 28 she founded a holography initiative at Perimeter. In 2026 she is building tools called WiPhy, PaperView and TheoryBench, and telling interviewers that Claude Code means a theorist no longer needs a dev team.* TRAINED IMSA (2010) · MIT (SB Physics, 2013) · Harvard (PhD, 2019, adviser Andrew Strominger) · Princeton PCTS (postdoc, 2019–22) AT Perimeter Institute (Research Faculty) · Simons Collaboration on Celestial Holography (Deputy Director) FILE N°074 Particle physics has a data problem. The Large Hadron Collider found the Higgs boson in 2012 and, as Natalie Wolchover put it in [Quanta Magazine](https://www.quantamagazine.org/is-particle-physics-dead-dying-or-just-hard-20260126/)in January 2026, nothing else showed up. Theorists have had to pick what to work on without experiments to steer them. Pasterski’s answer has been to go to the infrared instead: the lowest-energy part of scattering, the far edge of spacetime, and the symmetries that live there. Constraints in that region are rigid enough that you can make progress without new data. Now she is making a second bet: that the bottleneck in theoretical physics is partly an engineering problem. On this view, the corpus could be read, indexed and checked by machines if physicists build the tools themselves rather than leave it to AI labs. That is what Project Theoria is for. This file separates what she has documented from what has been said about her. Profile summary § 01 · The Build ## “Rivet, rivet, rivet, plane” Her father steered a childhood wish for a flying broomstick into flight lessons, and she took her first at age nine. Between 12 and 14 she built a single-engine Zenith CH 601 XL from a kit. The model had been involved in fatal in-flight breakups, so she made her own engineering modifications. Her father sold his car to pay for the kit and supervised the work. U.S. rules wouldn’t let a 14-year-old solo, but Canada would. On her 14th birthday she soloed a Cessna 150 at Sarnia, Ontario. That day a signed offer letter from Jeff Bezos for a future job at Blue Origin was waiting at home; she says it is technically still pending. Her [CV](https://physicsgirl.com/cv.pdf) lists an FAA Light Sport Aircraft Manufacturer credential (2008) and a U.S. solo permit (2009), and she flew her own plane solo at 16. She networked her way in. There were air shows and FAA officials, Krispy Kreme donuts brought to the FAA, and a photo book of the build carried through MIT hallways. Harvard rejected her and MIT put her on the waitlist. Press accounts say professors Allen Haggerty and Earll Murman saw her build video and advocated for her; she credits the networking. She got off the waitlist. The lesson she draws, on the Shawn Ryan Show: find the parts of a problem that are straightforward and systematic, and build something from them. She says she hasn’t fully lived that out yet. It is the thread that runs from the plane to Theoria. She left aerospace for three reasons, by her own account. - Internships at Blue Origin (2009), NASA Kennedy Space Center and Boeing Phantom Works (2011) showed her R&D that rarely shipped. - She worried about what could go wrong with suborbital products, including accidentally killing a rich customer. - Her aerospace heroes liked physics, so she switched to it. She now calls that the wrong reason. At MIT she won the inaugural Freshman Entrepreneurship Award (2011), did CERN research with the MIT-CMS group in the 2012 Higgs summer, and took her SB in 2013 with a 5.00 GPA and the Orloff Award. She notes that she has never founded a startup. § 02 · The Desert ## Why the infrared [Quanta’s first Qualia column](https://www.quantamagazine.org/is-particle-physics-dead-dying-or-just-hard-20260126/), published January 26, 2026, lays out the situation. The LHC has seen only Standard Model particles, and next-generation colliders face long timelines with no guaranteed discovery. Wolchover endorses the verdict of CERN theorist Cari Cesarotti, quoted below. Pasterski’s move follows from that. When experiments can’t reach high energies, look for exact low-energy statements that no ultraviolet completion can change. Soft theorems are statements of that kind. > “Particle physics isn’t dead; it’s just hard.” > Cari Cesarotti, CERN · Quanta Magazine, January 26 2026 § 03 · The Infrared Triangle ## Three corners, three links, and who did what The infrared triangle is the organizing picture of Andrew Strominger’s program, set out in his [Lectures on the Infrared Structure of Gravity and Gauge Theory](https://arxiv.org/abs/1703.05448) (arXiv:1703.05448). It has three corners, each discovered separately and decades apart. 1**Soft theorems**Weinberg (1965) Add one extremely low-energy photon or graviton to any scattering process. The amplitude factorizes into the original amplitude times a universal factor fixed by the external particles’ momenta and charges. 2**Asymptotic symmetries**Bondi, van der Burg, Metzner, Sachs (1962) At null infinity, where light rays end up, gravity has infinitely many symmetries. This is the BMS group. Supertranslations are direction-dependent time shifts; superrotations extend rotations and boosts. 3**Memory**Zel’dovich & Polnarev (1974); Christodoulou (1991); developed by Braginsky & Thorne A passing gravitational wave permanently displaces free test masses. The detector does not return to where it started. The sides of the triangle are the claims that each corner is the same fact as the others. - **Soft theorem ⇔ symmetry.** The soft theorem is the symmetry’s Ward identity, the constraint a symmetry places on amplitudes (Strominger and collaborators, 2013–14). - **Soft theorem ⇔ memory.** The memory formula is a Fourier transform in time of the soft theorem. Zero frequency means the long-time limit (Strominger & Zhiboedov, 2014). - **Symmetry ⇔ memory.** The wave carries the system between vacua related by a supertranslation, and memory records that vacuum transition. Pasterski’s part in it: - co-author of the 2014 subleading extensions: the superrotation, or Virasoro, symmetry of the S-matrix ([arXiv:1406.3312](https://arxiv.org/abs/1406.3312)) and Low’s subleading soft theorem as a QED symmetry (PRL 2014); - co-discoverer of spin memory (§ 04); - sole author of the electromagnetic triangle (§ 04); - author of the review *Implications of Superrotations* (Physics Reports, 2019); - and a user of the name: “The IR Triangle” has been one of her talk titles for years. Sidebar Stephen Hawking cited her spin-memory work before she finished her PhD, according to [Be Giant](https://www.begiant.ca/stories/people/sabrina-gonzalez-pasterski-perimeter-institute-physics). § 04 · Two Memories ## What is new and what is detectable **Spin memory** ([arXiv:1502.06120](https://arxiv.org/abs/1502.06120); JHEP 12 (2016) 053). Ordinary memory is a displacement caused by radiated energy. Spin memory is a time delay caused by radiated angular momentum: light beams on clockwise and counterclockwise orbits come back out of sync. Its formula is a Fourier transform of the subleading soft graviton theorem (Cachazo & Strominger, 2014), and its symmetry is the superrotation. Whether anyone can measure it is a separate question, and the honest answer is not soon. - It will not show up in single events. - David A. Nichols (Phys. Rev. D 95, 084048, 2017; arXiv:1702.03300) found that spin memory builds up during inspiral. He concluded it could be detected by carefully combining hundreds of compact-binary detections in future detectors such as the Einstein Telescope. - Grant & Nichols ([arXiv:2210.16266](https://arxiv.org/abs/2210.16266)) forecast an effective signal-to-noise ratio of 3 after 2 years with three Cosmic Explorer detectors, or 3.5 years with two. - They also note that spin memory is about ten times weaker than displacement memory, too weak for current ground-based detectors even across a population of events. - LISA is not the answer here. The same paper ties LISA’s memory prospects to displacement memory from supermassive black-hole mergers, not to spin memory. **Electromagnetic memory** ([arXiv:1505.00716](https://arxiv.org/abs/1505.00716); sole author; JHEP 09 (2017) 154). This paper completes the triangle for large U(1) symmetries at null infinity: a distant test charge gets a permanent velocity kick fixed by Weinberg’s soft-photon factor. The earlier classical analysis of electromagnetic memory is due to Bieri & Garfinkle (2013). Her own assessment, on the Shawn Ryan Show: she was lucky an observable came out of it, and it tests the boundary conditions more than it tests Einstein’s theory itself. § 05 · The Leaky Boundary ## Holography without a box AdS/CFT benefits from a timelike boundary that acts like a box, so the dual is an ordinary conformal field theory one dimension down. Our universe is not like that. At collider and gravitational-wave scales it is close to flat, and its boundary is null: radiation exits through future null infinity and never comes back. That is what “leaky” means here. Read u = t − r as a sentence *Retarded time is a distant observer’s clock reading minus the light-travel time.* Worked case: a flash leaves the origin at t = 0 and reaches r = 10 light-seconds at t = 10 s, so u = 0 everywhere on that wavefront. Null infinity is labeled by u plus a direction on the sky. Celestial holography proposes that the four-dimensional gravitational S-matrix equals correlators of a two-dimensional CFT on the celestial sphere, the sphere of sky directions. Pasterski offers two framings, and this file keeps both. In [Be Giant](https://www.begiant.ca/stories/people/sabrina-gonzalez-pasterski-perimeter-institute-physics)the image is that the universe casts a pattern on the inside of a giant sphere around you, and you read physics off it. On the Shawn Ryan Show she treats the equivalence as a useful reformulation, not a claim that we “are” a hologram. A reading-list note. The 2026 primer [Celestial Holography, A Hitchhiker’s Guide to the Celestial Sphere](https://arxiv.org/abs/2608.07568) (arXiv:2608.07568) is by Faïçal Barzi, not by her. § 06 · The Boost Basis ## The Mellin transform and conformal primaries The credit belongs to three papers: Pasterski, Shao & Strominger ([arXiv:1701.00049](https://arxiv.org/abs/1701.00049)); Pasterski & Shao ([arXiv:1705.01027](https://arxiv.org/abs/1705.01027)); and Pasterski, Shao & Strominger, *Gluon Amplitudes as 2d Conformal Correlators* ([arXiv:1706.03917](https://arxiv.org/abs/1706.03917)). The idea is this. The four-dimensional Lorentz group acts on the celestial sphere as SL(2,C), the global conformal group. So trade states of definite energy for states of definite boost weight Δ, which transform like two-dimensional CFT operators. The formula as a sentence φΔ(x; q) ∝ ∫0^∞ dω ω^Δ−1 e^∓iω q·x ∝ Γ(Δ) / (∓i q·x + ε)^Δ *Add up plane waves heading in sky direction q at every energy, weighting each by ω^Δ−1. The result depends on position only through q·x and rescales by a fixed power under boosts.* For massless particles, Δ = 1 + iλ gives a complete basis. Mini-example: with Δ = 1 the energy weighting is flat and the sum is 1/(q·x), a wave whose strength falls off away from the light ray in direction q. The result is that amplitudes become celestial correlators that transform like two-dimensional CFT correlators. This is the basis for her 2021 lectures and her 2023 chapter. § 07 · Soft, Then Hard ## Two lectures at the Institute for Advanced Study In July 2026 she gave a pair of lectures, *Asymptotic Symmetries and Detectors*, at the Institute for Advanced Study’s Prospects in Theoretical Physics program: Part 1 on July 22 and Part 2 on July 28. Both are in the player at the top of this page, and together they run just under two and a half hours. She closed the first with a map of the pair: “the goal today was soft, tomorrow is hard.” TL;DR - Part 1 is the infrared triangle done with observables. Einstein’s equations at null infinity constrain how the Bondi mass changes in time. Integrate that constraint and the charge splits in two: a soft part, which is the memory left in the waveform, and a hard part, which is the energy that matter and radiation carried out. - The memory seen in one direction of the sky is sourced by energy deposited in other directions. - The Ward identity is usually stated for the S-matrix, from an in-state to an out-state. She recasts it for in-in correlators, which are expectation values, the kind of quantity a detector reports. - Part 2 asks how universal the larger celestial symmetries are. She walks through three recent derivations, by other groups, that obtain the S algebra and w(1+∞) algebras from light-ray operators in conformal field theories. - Her verdict is cautious. These are operator algebras on a light sheet. She does not call the higher towers asymptotic symmetries in the standard sense, and she says the different derivations do not yet obviously connect. **Part 1 · Soft.** The timestamps below load the lecture at that moment. - She opens on the IR triangle of § 03 and gives her reason for returning to it: it is good to tie the formalism to physical observables, and to ask whether the newer celestial symmetries stand on the same footing as the old ones. - The constraint equations. The shear is free data, a function of time and sky direction. The Bondi mass and the angular-momentum aspect are not free; Einstein’s equations fix how they evolve. - The electromagnetic version, as intuition. An accelerated charge radiates, and knowing only its initial and final velocities already fixes the time integral of the radiated field. - The gravitational version. Two distant observers on geodesics oscillate as a wave passes and settle at a different separation. Integrating the geodesic deviation equation twice gives the memory. - A caveat she volunteers. This is the idealized measurement. A real one assumes the waveform from a single event is the only thing happening between source and detector. - The non-local picture. An energy detector at one point on the sky pings only when energy lands there. The soft graviton mode at that point is non-zero anyway, because it responds to energy sent in every other direction. - From the S-matrix Ward identity to in-in correlators, by inserting a complete set of states. Two-point functions of the soft charge become two-point functions of the energy operator. - The takeaway. The charges are supported on light rays: components of the stress tensor integrated along null infinity. That is common ground between celestial holography and the people who study light-ray operators and energy correlators. **Part 2 · Hard.** The second lecture turns from the soft charge to the hard one. - The problem. For the leading and subleading soft theorems, BMS symmetry gives a rigorous account. For the higher tower it is not obvious that every soft theorem has an asymptotic symmetry behind it, or that those theorems are universal. - The anchor result, due to Córdova and Shao: in a conformal field theory, light-ray operators built from the stress tensor on a light sheet realize the BMS algebra. - First derivation. A Kac–Moody symmetry plus conformal symmetry of the four-dimensional theory is enough to generate the S algebra, the gauge-theory counterpart of w. She flags the assumptions: tree level, linearized equations, and commutators that close on single-particle operators. - Second derivation. In a unitary four-dimensional CFT with a condition on its scalar spectrum, stress-tensor operators on a light ray can be sorted into celestial primaries that obey a w(1+∞) algebra. It rests on a result that only the identity and the stress tensor appear in stress-tensor commutators on a light sheet. - Third result, in brief. A claim that every three-dimensional CFT has a Λ-deformed w(1+∞) symmetry, with the CFT taken as the boundary dual of four-dimensional anti-de Sitter space. It is a different algebra reached in a similar spirit. - The question from the audience. Someone says they have seen half a dozen derivations of the w algebra that seem unrelated. She agrees it is confusing, and says she still sees a gap between the light-sheet derivations, the collinear-limit derivations, and anything that acts like a spacetime transformation. Neither lecture mentions AI. Any mapping from this material onto language models is a reader’s analogy, not hers. This summary was written from the videos’ auto-generated captions, which garble names and symbols. Check the video before quoting it. § 08 · Project Theoria ## Only what is documented Her CV dates Project Theoria to 2024. In [Be Giant](https://www.begiant.ca/stories/people/sabrina-gonzalez-pasterski-perimeter-institute-physics) (July 9, 2026) she describes building WiPhy, PaperView and TheoryBench so that AI and theoretical physics actually work together, instead of AI labs assuming they can solve physics without physicists. The status, as far as the record goes: - [Forbes Austria](https://www.forbes.at/artikel/das-universum-dekodieren-at)reports that Paperview was the first prototype, in April 2026; that a Unitree humanoid is meant to embody Theoria as a “purple paper eater”; and that the long-term target is postdoc-level capability. - Be Giant adds that she is buying a Unitree R1. - No public URLs or specifications exist for any of the three tools. On Claude Code she is specific. On the [Shawn Ryan Show](https://podcasts.apple.com/us/podcast/312-sabrina-pasterski-theoretical-physicist-on-the/id1492492083?i=1000772191414) (#312, published June 11, 2026) she says theorists mostly work with pen, paper and Mathematica, and that Claude Code lets one researcher do systematic computation that used to need hired developers. She rebuilt an earlier project with it within weeks. Her summary: “now I don’t need to hire a dev team.” In the same interview she defines “solving physics” as two jobs: axiomatize the laws, and curate and compress the corpus. She is wary of hype that pictures AI writing one paper that replaces researchers, and she wants industry and academia to collaborate on infrastructure. The institutional record: - Scientific Organizer, with Matthew Johnson and Anindita Maiti, of Perimeter’s [Theory + AI Symposium](https://events.perimeterinstitute.ca/event/951/)(April 7–8, 2025) and workshop (April 9–11, 2025); - “Ontology Czar” at Perimeter (2026); - teaching “AI Reasoning in Theoretical Physics” at Aspen (2026); - a planned “Theory of Theories” AI4Physics course (2027); - a talk at Argonne’s Genesis Mission (2026). \[reported\] She says she is using public attention to raise funding for Theoria. No funder or amount is public. Sidebar A · The Kaplan connection ## Two holographers from the same department, a decade apart Jared KaplanSabrina Pasterski IMSAKAPLAN · Class of 2001 · 2026 Alumni Trailblazer AwardPASTERSKI · Class of 2010 · 2019 Alumni Distinguished Leadership Award Harvard holographyKAPLAN · PhD 2009, “Aspects of Holography,” under Nima Arkani-HamedPASTERSKI · PhD 2019, under Andrew Strominger Theory + AI Symposium, April 7–8, 2025KAPLAN · Virtual keynote, “Human Level AI by 2030”PASTERSKI · Scientific Organizer, with Matthew Johnson and Anindita Maiti Path sinceKAPLAN · Joined OpenAI in 2019; co-founded Anthropic in 2021 (CSO); remains a Johns Hopkins associate professorPASTERSKI · Stayed in academia and builds AI tooling there; publicly credits Claude Code What is not documented: a co-authored paper, an Anthropic–Theoria partnership, or any on-record conversation between the two. The [Kaplan file](https://www.contextjamming.com/founder-files/kaplan) is N°006. Sidebar B · Kaplan’s Quanta prediction In the same [Quanta column](https://www.quantamagazine.org/is-particle-physics-dead-dying-or-just-hard-20260126/) (January 26, 2026), Kaplan makes four predictions. In paraphrase: - AI will progress faster than almost any field of science has historically. - Planning on a 10-year horizon barely matters, because any accelerator built in a decade would be built by AI. - There are roughly even odds that theoretical physicists are mostly replaced by AI within two or three years. - AI may soon write theory papers at the level of Nima Arkani-Hamed or Ed Witten. Cesarotti, in the same column, worries that AI is making people worse at physics. Pasterski’s documented view sits between the two: AI tools transform theory work, but physicists must stay in the loop. § 09 · Can Computers Understand? ## A debate with no transcript The event was the IAI / HowTheLightGetsIn London debate [“Cracking the Code for Thought,”](https://iai.tv/video/cracking-the-code-for-thought)with Roger Penrose and Max Tegmark (by Zoom), hosted by Güneş Taylor. Her CV dates it September 20, 2025. A clip circulates under the title “Can computers understand?” Penrose’s position is well known: understanding is non-computable, by a Gödel argument, with the Orch-OR proposal he developed with Hameroff. Pasterski’s on-stage argument is not documented in an accessible transcript, so this file does not characterize it. The closest documented view comes from the Shawn Ryan Show: physics doesn’t obviously rule out neural-net-like systems making decisions, and she tentatively favors treating the observer as part of the physical system. Two related events are on the record as events only: an AGI debate with Sergey Brin (Aspen, January 2026) and an AI debate with Neil Turok (London, September 20, 2026). § 10 · Institutions ## Perimeter, the initiative, and the collaboration **Perimeter.**She joined the faculty in 2021 after PCTS and moved to Waterloo in 2022. She values Perimeter’s low-noise, theory-only design and Canada’s relatively low prestige-chasing, as she tells Be Giant. **Celestial Holography Initiative (2021).** She is the principal investigator, leading researchers in amplitudes, mathematical physics and quantum gravity. Recent co-authors include Yangrui Hu and Adam Ball. **Simons Collaboration on Celestial Holography.** Announced August 21, 2023 and directed by Strominger; she is a deputy director. Her CV lists $7,995,000 shared (2023) and a Simons intent to fund from September 2027 to August 2030. **Strings 2023.** Lead Coordinator. \[reported\] The Shawn Ryan Show introduction describes her as the first woman to chair Strings. Her work is broadening into quantum information and holography. Recent titles: *Radiation in Holography*, *Cryptographic Tests of the Python’s Lunch Conjecture*, *Swing Surfaces in AdS/CFT*, and *Generalized Entanglement Wedges and the Connected Wedge Theorem*. § 11 · The Attention Paradox ## The label, the plane, and the lectures She traces the “next Einstein” label to an Ozy Media story during the plane build. She never asked for it and has always been wary of it. On the Shawn Ryan Show she called it inaccurate and said she probably benefits from it too much. She also observed that the physicists the public knows are often not the ones the field rates most highly. The recognition is real all the same: Scientific American 30 Under 30 (2012); Forbes 30 Under 30 (2015) and All-Star (2017); and the Met Gala in May 2026, in a dress she 3D-printed from Gemini-assisted designs based on the Hopf fibration. The paradox is that the public story is the plane and the label, while the work, such as the 2021 celestial-amplitudes lectures, reaches mostly specialists. > “I’m treating the spotlight as a resource, not a trophy.” > Pasterski, Be Giant, July 9 2026 The Index 14 Age at first solo — Cessna 150, Sarnia CH 601 XL Zenith kit aircraft, built at ages 12–14 5.00 MIT GPA, out of 5.00 — Orloff Award 2013 3 Triangle corners — soft theorem · symmetry · memory 4D → 2D Bulk to boundary — the celestial sphere $7,995,000 Simons funding, shared, 2023 — per her CV 3 Theoria tools named — WiPhy · PaperView · TheoryBench ~50% Kaplan’s odds that AI mostly replaces theorists in 2–3 years — Quanta 2026 ~10× How much weaker spin memory is than displacement memory; detection needs stacked events in 3G detectors — Grant & Nichols forecast, arXiv:2210.16266 Timeline - 1993Born June 3 in Chicago. - 2003First flight lesson, at age nine. - 2005Founds Sabrina Aircraft Manufacturing (per her CV). - 2005–07Builds the Zenith CH 601 XL from a kit. - 2007Solos a Cessna 150 at Sarnia, Ontario, on her 14th birthday. The Bezos letter arrives the same day. - 2008FAA Light Sport Aircraft Manufacturer credential. - 2009Blue Origin intern. U.S. solo permit. - 2010Graduates from IMSA. - 2011MIT’s inaugural Freshman Entrepreneurship Award. Internships at NASA Kennedy Space Center and Boeing Phantom Works. - 2012CERN, with the MIT-CMS group, during the Higgs summer. Scientific American 30 Under 30. - 2013MIT SB in physics with a 5.00 GPA. Orloff Award. - 2014The superrotation and Low’s-theorem papers. - 2015Spin memory. Electromagnetic memory. Forbes 30 Under 30. - 2017The conformal-basis papers. - 2019Harvard PhD under Andrew Strominger. “Implications of Superrotations” in Physics Reports. Starts at Princeton’s PCTS. - 2021Joins Perimeter. Founds the Celestial Holography Initiative. Lectures on Celestial Amplitudes. Snowmass white paper. - 2023Deputy director of the Simons Collaboration on Celestial Holography. Lead Coordinator of Strings 2023. Founders & Physicists (Canada). - 2024Founds Project Theoria (per her CV). - 2025Theory + AI Symposium at Perimeter (April). The Penrose debate in London (September). - 2026Brin debate at Aspen (January). Paperview prototype (April). Met Gala (May). Shawn Ryan Show #312 (June 11). Be Giant profile (July 9). Two lectures at the Institute for Advanced Study (July 22 and 28). Turok debate in London (September). Key works - [2014Semiclassical Virasoro Symmetry of the Quantum Gravity S-MatrixarXiv:1406.3312 · with Kapec, Lysov, Strominger →](https://arxiv.org/abs/1406.3312) - [2015New Gravitational MemoriesarXiv:1502.06120 · with Strominger, Zhiboedov →](https://arxiv.org/abs/1502.06120) - [2015Asymptotic Symmetries and Electromagnetic MemoryarXiv:1505.00716 · sole author →](https://arxiv.org/abs/1505.00716) - [2017Flat Space Amplitudes and Conformal Symmetry of the Celestial SpherearXiv:1701.00049 · with Shao, Strominger →](https://arxiv.org/abs/1701.00049) - [2017A Conformal Basis for Flat Space AmplitudesarXiv:1705.01027 · with Shao →](https://arxiv.org/abs/1705.01027) - [2019Implications of SuperrotationsarXiv:1905.10052 · Physics Reports →](https://arxiv.org/abs/1905.10052) - [2021Lectures on Celestial AmplitudesarXiv:2108.04801 · EPJC →](https://arxiv.org/abs/2108.04801) - [2021Celestial HolographyarXiv:2111.11392 · Snowmass · with Pate, Raclariu →](https://arxiv.org/abs/2111.11392) - [2023A Chapter on Celestial HolographyarXiv:2310.04932 · Elsevier 2025 →](https://arxiv.org/abs/2310.04932) - [ContextLectures on the Infrared Structure of Gravity and Gauge TheoryarXiv:1703.05448 · Andrew Strominger →](https://arxiv.org/abs/1703.05448) - [ContextCelestial Holography, A Hitchhiker’s Guide to the Celestial SpherearXiv:2608.07568 · Faïçal Barzi · not hers →](https://arxiv.org/abs/2608.07568) Sources 1. 01[Pasterski — curriculum vitae](https://physicsgirl.com/cv.pdf) 2. 02[Perimeter Institute — faculty page](https://perimeterinstitute.ca/people/sabrina-pasterski) 3. 03[Simons Collaboration on Celestial Holography — people](https://simonscelestialholographycollaboration.org/people/) 4. 04[Be Giant — profile, July 9, 2026](https://www.begiant.ca/stories/people/sabrina-gonzalez-pasterski-perimeter-institute-physics) 5. 05[Forbes Austria — “Das Universum dekodieren”](https://www.forbes.at/artikel/das-universum-dekodieren-at) 6. 06[Shawn Ryan Show #312 — Apple Podcasts, June 11, 2026](https://podcasts.apple.com/us/podcast/312-sabrina-pasterski-theoretical-physicist-on-the/id1492492083?i=1000772191414) 7. 07[Shawn Ryan Show #312 — transcript](https://podcasts.happyscribe.com/the-shawn-ryan-show/312-sabrina-pasterski-theoretical-physicist-on-the-hidden-code-of-the-universe) 8. 08[Perimeter — Theory + AI Symposium](https://events.perimeterinstitute.ca/event/951/) 9. 09[Perimeter — Theory + AI Symposium agenda](https://events.perimeterinstitute.ca/event/951/page/449-agenda) 10. 10[PIRSA — Theory + AI collection](https://pirsa.org/c25020) 11. 11[Quanta Magazine — “Is Particle Physics Dead, Dying, or Just Hard?”](https://www.quantamagazine.org/is-particle-physics-dead-dying-or-just-hard-20260126/) 12. 12[IMSA — 2026 Board of Trustees Alumni Awards](https://www.imsa.edu/2026-imsa-board-of-trustees-alumni-awards/) 13. 13[Wikipedia — Jared Kaplan](https://en.wikipedia.org/wiki/Jared_Kaplan) 14. 14[IAI — “Cracking the Code for Thought”](https://iai.tv/video/cracking-the-code-for-thought) 15. 15[arXiv:1502.06120 — New Gravitational Memories](https://arxiv.org/abs/1502.06120) 16. 16[arXiv:1505.00716 — Asymptotic Symmetries and Electromagnetic Memory](https://arxiv.org/abs/1505.00716) 17. 17[arXiv:2310.04932 — A Chapter on Celestial Holography](https://arxiv.org/abs/2310.04932) 18. 18[arXiv:2608.07568 — Barzi, A Hitchhiker’s Guide to the Celestial Sphere](https://arxiv.org/abs/2608.07568) 19. 19[arXiv:2210.16266 — Grant & Nichols, spin-memory detection forecast](https://arxiv.org/abs/2210.16266) 20. 20[IAS — Asymptotic Symmetries and Detectors, Part 1 (video, July 22, 2026)](https://www.youtube.com/watch?v=t3G2FmoJYdE) 21. 21[IAS — Asymptotic Symmetries and Detectors, Part 2 (video, July 28, 2026)](https://www.youtube.com/watch?v=hvHaGL3hPww) 22. 22[SlideRoom — “How Admissions Missed the Next Einstein”](https://www.slideroom.com/blog/how-admissions-missed-the-next-einstein/) Related reading [Jared Kaplan, PhD · FounderFiles N°006](https://www.contextjamming.com/founder-files/kaplan) The other IMSA and Harvard holographer, who left for the labs [Ethan Dyer, PhD · FounderFiles N°061](https://www.contextjamming.com/founder-files/dyer) A holographer who wrote the theory of why scaling laws exist [Juan Maldacena · FounderFiles N°009](https://www.contextjamming.com/founder-files/maldacena) AdS/CFT, the holography whose boundary acts like a box [The Data Desert: Physics, AI, and the Diaspora](https://www.contextjamming.com/PhysicsAI-Diaspora) Context Jamming · why physicists crossed into AI, and who stayed Career Shape I-shaped — a single maximal-depth spike ### I-Beam Theorist Drives one domain to maximal depth and lets the world reorganize around the result; commercialization is downstream, optional, or never. Role-Model Reference Class - Andrew Strominger - Stephen Hawking - Roger Penrose Founder Context · JSON A small reasoning persona distilled from this file. Inject it into a chat or deep-research context to assess a business problem the way Pasterski would. Reason as a theorist of the infrared. Find the systematic, buildable part of a hard problem and start there. Prefer exact, symmetry-protected statements at the boundary over claims that depend on unknown high-energy detail. Ask whether a proposed observable tests the theory or only its boundary conditions. Treat the existing body of knowledge as something to curate and compress, and keep the practitioner in the loop when a machine does the systematic work. ``` { "$schema": "https://www.contextjamming.com/schemas/founder-context-v1.json", "file": "N°074", "persona": "Dr. Sabrina Gonzalez Pasterski", "archetype": "i-beam", "shape": "I", "one_line": "A theorist of the infrared who looks for the systematic, buildable part of a hard problem, in physics and now in AI tooling for physics.", "cognitive_basis": { "credentialPath": "doctoral", "abstractionDirection": "top-down", "exitHorizon": "deferred", "moatInstinct": "theoretical-insight", "capitalPosture": "not-established" }, "operating_questions": [ "Which parts of this problem are straightforward and systematic enough to build from?", "What exact statement holds at low energy, or at the boundary, no matter what the unknown high-energy completion turns out to be?", "Is there a symmetry behind this constraint, and what does it force?", "Could t … ``` Share FounderFiles N°074 · Sabrina Gonzalez Pasterski, PhD Filed by Bret Kerr · ACRA Insight LLC · Franklin, MA contextjamming.com · @bretkerr [← back to Context Jamming](https://www.contextjamming.com/)