{
    "componentChunkName": "component---src-templates-article-js",
    "path": "/article/2026/05/31/making-light-work-of-matrix-multiplication",
    "result": {"data":{"strapiArticle":{"id":"Article_6a1b55f9a1914505ee622e86","title":"Making light work of matrix multiplication","subtitle":"Efficient computing to speed up AI","published_at":"2026-05-31T19:32:07.608Z","content":"Meeting the massive energy\ndemands of the AI boom may\nrequire an outside-the-box solution. Some researchers believe the answer\nlies in light.\n\nThe idea of computing with light\nrather than electricity has been around\nsince the 1960s, but physicists and engineers have recently renewed their efforts\nin this area, hoping to decrease the energy\nrequired to execute machine learning programs. Building a light-based or “optical”\ncomputer requires precise engineering\nthat has taken decades to develop, but\nthe basic principle behind such devices\nis something you’ve seen in grade school.\n\nThink of that time you put a pencil\nin a cup of water and the pencil looked\nbroken. It looked broken because air and\nwater interact with light differently. Put\nin scientific terms, the air and water have\ndifferent refractive indices, and any time\na wave of light encounters a change in\nrefractive index, it will bend. Therefore,\none can precisely manipulate the path\nof light through a material by controlling the refractive index. In particular,\nif one could change the refractive index\nof a material independently at specific\nlocations, an “input” wave of light would\nbend differently at each location, creating\na completely new “output” wave.\n\nThe bending of an “input” wave into a distinct “output” wave is one example\nof a light-based computation. It is, in\nfact, the easiest type of computation\nto perform with light. But the reason\nlight-based computing has gotten so\nmuch attention over the past few years is\nthat, mathematically, this bending can\nrepresent the most important computation for machine learning: matrix-vector\nmultiplication.\n\nA matrix is a table of numbers, and\na vector is a column of numbers. Multiplying a matrix and a vector involves\nmultiplying every number in the table by\nevery number in the column and adding\nup the results to make a new vector. The\nvectors that ChatGPT uses are thousands\nof numbers long, and to produce every\nnew word it writes, it must perform\nhundreds of matrix-vector multiplications. With millions of users asking the\nchatbot to write thousands of words per\nday, you can quickly see how increasing\nthe energetic efficiency of matrix-vector\nmultiplication is critical.\n\nThe light-based computing device\nI described above is one way to implement matrix-vector multiplication with\nlight, and it’s roughly equivalent to one\nthat Professor Peter McMahon and his\ncolleagues designed last year. The team,\nbased at Cornell University, published\ntheir findings in Nature Physics last December. In their device, the matrix is\nphysically encoded as a grid in a slab of\nspecial material, where each cell has a\ndifferent refractive index. The grid is created by a temporary and programmable\nprocess, allowing the same material to\nrepresent any matrix.\n\nOne way to understand the computational ability of their device is through\na geometric interpretation of matrices and\nvectors. You can consider a vector not as a\ncolumn of numbers but as an arrow pointing in some direction. Multiplying by a\nmatrix rotates and stretches that arrow\nto form a new arrow with a different size\nand direction. Similarly, the wave of light\ntraveling across the device can be viewed\nas an arrow in a high-dimensional space.\nThe changes in refractive index across the\nmaterial bend the wave of light at every\nlocation and essentially act as one big,\ncomplicated rotation. By tuning the grid\nof refractive indices, one can replicate any\ntype of rotation and therefore any matrix.\n\nThe geometric interpretation of\nmatrix-vector multiplication also reveals\nwhy a light-based approach can be more\nenergetically efficient than an electronic\napproach. The bending of light in\nresponse to a change in refractive index\nis a natural phenomenon. Once the\ndevice is made, additional energy input\nis not required for rotation. In contrast,\na traditional electronic computer would\nsimply multiply and add a series of numbers. Each numerical operation requires\ncharging and discharging a set of wires,\ncarrying with it an intrinsic energy cost.\n\n![Oaks-LeafBlackburnThumbSpring2026.jpg](https://www.berkeleysciencereview.com/api/uploads/Oaks_Leaf_Blackburn_Thumb_Spring2026_feacafadf5.jpg)\n\nProfessor McMahon explains that\nthe propagation of light through a material “can be, in principle, completely\nlossless,” meaning that no energy is lost\nduring the computation. Some energy\nwill be dissipated in an optical matrix-vector multiplication from the light being\nabsorbed by the device, but it’s negligible\ncompared to the energetic cost of electronic computations. It’s for mainly this\nreason that 30 years ago there was already\na major shift away from electronics and\ntowards optics in the field of communication. Today, almost all long-distance communication is done with optical signals ([https://www.noaa.gov/submarine-cables](https://www.noaa.gov/submarine-cables)) because much less energy is lost when\nlight travels through a fiber-optic cable\nthan when electricity travels through a\nwire. Computing, however, is more complicated than communication. Several\ntechnological challenges could prevent\noptical computing from proliferating as\nquickly as optical communication once\ndid.\n\nOne major concern is that it requires\na significant amount of energy to convert light into an electrical signal and\nvice versa. This would be necessary if\nthe optical computing device performed\nonly matrix-vector multiplication within\na larger program, which many researchers view as the most feasible near-term\nuse for optical computing. Even with\nthis conversion cost, optical computing\ncould still be more energetically efficient\nthan electronic computing, but crucially\nonly at a certain size of input vector. For\nsmaller vectors, the conversion cost outweighs the benefits. Luckily, modern AI\nalgorithms do use massive vectors and\nmatrices—large enough to almost surely\ndwarf the conversion cost. But that brings us to our next practical challenge: size.\n\nA typical wavelength of light used\nin optical computing is more than 100\ntimes larger than a modern transistor.\nThat wavelength sets the scale for the\nsize of the optical device, meaning it is\ndifficult to make optical processors as\nsmall as their electronic counterparts. On\nthe millimeter scale, McMahon’s team\ndemonstrated that their device could\nprocess input vectors of length 49—the\nlargest demonstration in a device of this\nkind. Some researchers are optimistic that\nmillimeter-scale devices could one day\nhandle input vectors near the size used\nin modern AI programs, but it has yet to\nbe shown experimentally.\n\nTo make large-scale optical devices\nfeasible, researchers utilize another key\nadvantage of optics: the high clock rate.\nClock rate is the term computer scientists\nuse to describe the number of mathematical operations a computer performs per\nsecond. The charging and discharging\nof wires in a digital computer not only\ndissipates heat but also puts a limit on\nthe clock rate. In fact, clock rate and\nheat dissipation are linked: the faster\nyou charge and discharge the wires in\nan electronic computer, the more energy\nit will use. Optical computing, however,\ndoesn’t have this problem. One can easily\nachieve a clock rate 10–100 times faster\nthan a digital computer without dissipating significant energy.\n\nThis clock rate advantage is the key\nidea behind the startup Opticore, founded\nby UC Berkeley professors Zaijun Chen\nand Mengjie Yu along with MIT research\nscientist Ryan Hamerly. They believe that\noptical processors will be used in large-scale AI platforms in the near future.\nTheir technology relies on using time as\none of the dimensions of the matrix. For\nexample, in one of Opticore’s devices,\nthe part of the processor that represents\nthe matrix changes over time. Thus, they\nessentially perform a large matrix-vector\nmultiplication by breaking it up into several smaller multiplications, all executed\nin extremely quick succession.\n\nProfessor Chen predicts that optical processors will be used in large scale\napplications “within five to ten years.” He\nnotes that major companies like Nvidia\nare already beginning to integrate optics\ninto their computer chips to transport\ndata. He feels that this will pave the way\nfor light not only to move the data but\nto manipulate it. Other researchers in the\nfield, however, are more cautious. Martin Stein, a postdoctoral researcher working on optical computing with Professor\nMcMahon at Cornell University, points\nout that “many of the smartest and best-paid people in the world are trying to\nmake machine learning more energy efficient with the hardware that’s currently available, and they are advancing with\nan incredible pace that is very difficult\nfor us to keep step with.” His concerns\nare rooted in the fact that no one has yet\ndemonstrated an optical matrix-vector\nmultiplier large enough to be practical.\nAnd even if one were built, it would still\nbe extremely costly to alter existing architectures to integrate it.\n\nIf widely implemented in AI, optical computing could cause a substantial\nparadigm shift in technology, perhaps\nmost akin to the widespread adoption\nof the optical cables that now power the\ninternet ([https://www.submarinecablemap.com](https://www.submarinecablemap.com)). But, as with the early stages of\nany new technology, there are both doubters and true believers in the optical computing field. It is likely too early to tell\nwho will ultimately be right, but one can\nonly hope that the specter of skyrocketing\nenergy costs may be enough incentive for\nboth researchers and industrial leaders to\nmake some kind of creative computing\nsolution work.","image":{"publicURL":"/static/f63d20e26e94ffcca4c42deee4a15a82/8c12ed5e08b656d692f5dc0da7928d45.png"},"authors":[{"id":"638bfcad1e0c2a0a6f337335","name":"Sam Oaks-Leaf"}],"designers":[{"id":"692fe677a1914505ee622dda","name":"Albany Blackburn"}],"categories":[{"id":"6030b07d7782326a48b058be","title":"Tech & AI"}],"magazine":{"id":"6a1b3daba1914505ee622e6e","title":"Spring 2026","issue":50}},"recent":{"edges":[{"node":{"id":"Article_6a1b7b70a1914505ee622e8a","title":"A quarter-century of BSR","authors":[{"id":"6930d4c5a1914505ee622e05","name":"Eleanor Wang"}],"categories":[],"image":{"publicURL":"/static/e2535987c5657a0ccb90aefc70d390fa/d0381637f7e57d0c526bc22c6b39f421.png"},"published_at":"2026-05-31T19:32:44.843Z"}},{"node":{"id":"Article_6a1c7dfda1914505ee622e93","title":"A taste of Ohlone culture","authors":[{"id":"6a1b3e7ea1914505ee622e74","name":"Jack Schill"}],"categories":[{"id":"61fdc5cd1e0c2a0a6f3372c1","title":"Science and Society"}],"image":{"publicURL":"/static/7c0a8d7857085f221c3cf6c5b6578d0e/4d0dc80edda1095b8e6507ebeafd9a81.jpg"},"published_at":"2026-05-31T19:32:37.016Z"}},{"node":{"id":"Article_6a1c7d2ea1914505ee622e91","title":"The rains are coming","authors":[{"id":"684f9826a1914505ee622dcb","name":"Savvas Marcou"}],"categories":[{"id":"6030b07d7782326a48b058bc","title":"Climate Change"},{"id":"61fdc5cd1e0c2a0a6f3372c1","title":"Science and Society"}],"image":{"publicURL":"/static/f8a4923043b054cd1e972c4da212c28e/c5850deda96224cd2379429077ac4146.jpg"},"published_at":"2026-05-31T19:32:26.544Z"}},{"node":{"id":"Article_6a1b55f9a1914505ee622e86","title":"Making light work of matrix multiplication","authors":[{"id":"638bfcad1e0c2a0a6f337335","name":"Sam Oaks-Leaf"}],"categories":[{"id":"6030b07d7782326a48b058be","title":"Tech & AI"}],"image":{"publicURL":"/static/f63d20e26e94ffcca4c42deee4a15a82/8c12ed5e08b656d692f5dc0da7928d45.png"},"published_at":"2026-05-31T19:32:07.608Z"}}]},"related":{"edges":[{"node":{"id":"Article_6a1b55f9a1914505ee622e86","title":"Making light work of matrix multiplication","authors":[{"id":"638bfcad1e0c2a0a6f337335","name":"Sam Oaks-Leaf"}],"categories":[{"id":"6030b07d7782326a48b058be","title":"Tech & AI"}],"image":{"publicURL":"/static/f63d20e26e94ffcca4c42deee4a15a82/8c12ed5e08b656d692f5dc0da7928d45.png"},"published_at":"2026-05-31T19:32:07.608Z"}},{"node":{"id":"Article_6a1b50c9a1914505ee622e83","title":"Food for bots","authors":[{"id":"6a1b3e1da1914505ee622e73","name":"Albany Blackburn"}],"categories":[{"id":"6030b07d7782326a48b058be","title":"Tech & AI"},{"id":"61fdc5cd1e0c2a0a6f3372c1","title":"Science and Society"}],"image":{"publicURL":"/static/01a5a957d327939696b73dbbbfdad91d/9b64e41f908c29b3f30d49ecc1066119.png"},"published_at":"2026-05-31T19:30:49.523Z"}},{"node":{"id":"Article_6930d9a2a1914505ee622e10","title":"The Redwood Center gets its twentieth ring","authors":[{"id":"645880ea1e0c2a0a6f3373b4","name":"Deepthi Bannai"},{"id":"692fe8a5a1914505ee622de3","name":"Galen Chuang"}],"categories":[{"id":"6030b07d7782326a48b058bd","title":"Life Science"},{"id":"6030b07d7782326a48b058be","title":"Tech & AI"}],"image":{"publicURL":"/static/27f82cb285643ccc060ef0d6a28fdde7/bd7f1fe76f18fc8605e4e9baa6453766.jpg"},"published_at":"2025-12-04T01:39:28.955Z"}},{"node":{"id":"Article_681ea29354c55417f913848e","title":"Leading the charge to better batteries","authors":[{"id":"681e8cdf54c55417f9138472","name":"Sydney Hemenway"}],"categories":[{"id":"6030b07d7782326a48b058be","title":"Tech & AI"}],"image":{"publicURL":"/static/31a6b0953c8f5dfcfac0633d6188c1e8/d53b7694f86e6625642e943fab44816d.jpg"},"published_at":"2025-05-10T00:57:32.172Z"}}]},"allStrapiMagazineIssue":{"edges":[{"node":{"id":"Magazine-issue_6a1b3daba1914505ee622e6e","title":"Spring 2026","pdf":{"publicURL":"/static/141d2a4c7f0cefde02f65106eeb86d9a/61ab272b58a1f4df0a197f05dd52fcf6.pdf"}}},{"node":{"id":"Magazine-issue_692fe437a1914505ee622dd4","title":"Fall 2025","pdf":{"publicURL":"/static/fac3533f3374e2104f7af13a628ac32b/3044818f0fe1c2efaa51629f83bd8fec.pdf"}}},{"node":{"id":"Magazine-issue_68438450a1914505ee622dc5","title":"Spring 2025","pdf":{"publicURL":"/static/5a530985b7bc153f8ba2603b00e5fad0/4d32772a90aaea4dd24116a9bb228d7f.pdf"}}},{"node":{"id":"Magazine-issue_674cca8e54c55417f913842b","title":"Fall 2024","pdf":{"publicURL":"/static/3a00f64ea9907fd8a3c34f2532f44d84/932bda9a0de968d6d83633dae23f54e6.pdf"}}},{"node":{"id":"Magazine-issue_66b45edb1d44e161e6a9acfb","title":"Spring 2024","pdf":{"publicURL":"/static/e3536b6f0c7152212f93396babf7c624/9d224496ebfd458878d2ea9e43d94d02.pdf"}}},{"node":{"id":"Magazine-issue_66b45d801d44e161e6a9acf8","title":"Fall 2023","pdf":{"publicURL":"/static/5940a45251e7ea7ef5bc6e33fec32faf/2ef8c8e8e651cb8f83db9a992c881ba7.pdf"}}},{"node":{"id":"Magazine-issue_646d00c81e0c2a0a6f3373c0","title":"Spring 2023","pdf":{"publicURL":"/static/9ced8e2bf26e9ee5985fe0fc63451ef3/27617aed1b53bc874575b0c4faef49ed.pdf"}}},{"node":{"id":"Magazine-issue_638c146d1e0c2a0a6f33735c","title":"Fall 2022","pdf":{"publicURL":"/static/4eb6d78b1756c94dded828c15d583c4c/aa08ea0bd364f78674f4f1a6d4d6abe0.pdf"}}},{"node":{"id":"Magazine-issue_626e16711e0c2a0a6f3372dd","title":"Spring 2022","pdf":{"publicURL":"/static/ac176ebf6c3c4dfb4641afe166325b18/640512489214c8aff2868a8f9a4386ea.pdf"}}},{"node":{"id":"Magazine-issue_619d303c1e0c2a0a6f337222","title":"Fall 2021","pdf":{"publicURL":"/static/5e4c68b3d0095649c7644a0c1389e259/98982c48adcab02a7902633e3c020aed.pdf"}}},{"node":{"id":"Magazine-issue_609466b619a23cc0de32887b","title":"Spring 2021","pdf":{"publicURL":"/static/62802cea357974db7ac5d7f7180f4f9f/b2c551b065cccb83f1cd3cc7b1c6c0a2.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058da","title":"Fall 2020","pdf":{"publicURL":"/static/022b8deca25507f2a713dd583ceee964/8dcfdaff0645e6fcaeae1ad88e2de9b7.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d9","title":"Spring 2020","pdf":{"publicURL":"/static/ab06dd12e4becd100db5bec830483b31/b5331d5386b9d845c70d1882a473028e.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d4","title":"Fall 2019","pdf":{"publicURL":"/static/df36dae2b996ff04ee930a334e15ab4c/db397e737ecaf2062193bc0884fa49bc.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d8","title":"Spring 2019","pdf":{"publicURL":"/static/552e5aaa76ed60792d3472c9226d849d/d5f91b2dc8312c7cd2d17da2ad99e571.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d7","title":"Fall 2018","pdf":{"publicURL":"/static/0413276edd1544f5113278ca7f504a5d/bda97b47b7fab30498bb6f939c57a30f.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d6","title":"Spring 2018","pdf":{"publicURL":"/static/78a3086d4ba2dcfac6980388801ad170/c3dfdbf4de030acab92f6067f8581f06.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d5","title":"Fall 2017","pdf":{"publicURL":"/static/3cec335d3d6e3583b33b183e7afef1ef/252744cc30b7ee760bc1c82da05d5d22.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d3","title":"Spring 2017","pdf":{"publicURL":"/static/3a8d3f95f7c0e412cae53c6f2a359bcc/97888accf55a6567b2099a487e8d728c.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d2","title":"Fall 2016","pdf":{"publicURL":"/static/36d42f5373e261271870dbf8ca637ec5/4f82cf8088cdb7ceb472a6df293a403f.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d1","title":"Spring 2016","pdf":{"publicURL":"/static/bc7e5ea86f4d73abdcd1e5da5fbb97dc/b39d9d2f9c6e5d3e9c087198857e7b1e.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058d0","title":"Fall 2015","pdf":{"publicURL":"/static/77d59b254f897d4e092af2ec9186be0a/57ff22c4b4e09cab75c1e2d8fce15d8f.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058cf","title":"Spring 2015","pdf":{"publicURL":"/static/dd45dc3f97cbe6eb2c2af82975446e81/ef40b63b0e330f3dc9c1fd4869656668.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058ce","title":"Fall 2014","pdf":{"publicURL":"/static/c1bf635b72dcec6e54cd8278591699cd/e129bcf90df0e9bdc6b9ec91daf211fd.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058cd","title":"Spring 2014","pdf":{"publicURL":"/static/b8cb5e598290f48b81efa678ea0b8f3f/e2549a5110876096fc7d043b0259a0bc.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058cc","title":"Fall 2013","pdf":{"publicURL":"/static/8e5d6661e26b49667cf83b4fffbc0bf5/9f385169503fb1df2d6eca4b1a94b078.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058cb","title":"Spring 2013","pdf":null}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058c9","title":"Fall 2012","pdf":{"publicURL":"/static/d46c899fbddd6e61c58dc2a642f8fcf6/32f3565e40340fc397e7fd1fd3a5495c.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058ca","title":"Spring 2012","pdf":{"publicURL":"/static/5fb470b87607d223a2da0eee1810ffbb/6deb8076010cba9bb125e04f72c1b99c.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058c7","title":"Fall 2011","pdf":{"publicURL":"/static/bdca8874973420fddae3d4b21e6cd0c7/b426ddfb9ef21c3e4eca5a88e71e833b.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058c8","title":"Spring 2011","pdf":{"publicURL":"/static/2afa203fc148a40b4f35542883915ef7/7d1e487d18e504497d8d7cb01ec2234c.pdf"}}},{"node":{"id":"Magazine-issue_69840835a1914505ee622e27","title":"Fall 2010","pdf":{"publicURL":"/static/19f3b3aae3a55bd1903646c9f09f796d/c5d776d7e0ed1b9f22aee55fcbea6dec.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058c6","title":"Spring 2010","pdf":{"publicURL":"/static/e47ff4ef74bf8dda8fd9a8c19f53d2c7/2c0e2ea73d8e66f566b3d5738cdbeec1.pdf"}}},{"node":{"id":"Magazine-issue_61b7856c1e0c2a0a6f33728c","title":"Fall 2009","pdf":{"publicURL":"/static/e282cf0df70cc3ac58b3479918f3f701/3679aac8933b440018668573eb07f64d.pdf"}}},{"node":{"id":"Magazine-issue_61b786981e0c2a0a6f337291","title":"Spring 2009","pdf":{"publicURL":"/static/9656dcfa953cd604c9738d0846e4e708/74296aaaafa9c07043caa2a043c30d13.pdf"}}},{"node":{"id":"Magazine-issue_69841dd6a1914505ee622e2c","title":"Fall 2008","pdf":{"publicURL":"/static/8836a4c0f4834b01c65b2185860dbca7/538dcb7fc3186af282b095d0981fe339.pdf"}}},{"node":{"id":"Magazine-issue_61b7870b1e0c2a0a6f337298","title":"Fall 2007","pdf":{"publicURL":"/static/0fa4cf6d85148e7a8ec1f8a2fb816e8e/5e20d89d6e81f5dfd0d3da1e12ba0f35.pdf"}}},{"node":{"id":"Magazine-issue_698423a7a1914505ee622e4f","title":"Fall 2006","pdf":{"publicURL":"/static/82cb80f4b0112a111d639959eb26e90c/458ae0f8181ed899c5bf4bbb329093fb.pdf"}}},{"node":{"id":"Magazine-issue_69842222a1914505ee622e40","title":"Spring 2006","pdf":{"publicURL":"/static/c10933b8c9558129ee4658ef89ce2cc2/d800afdd672e519a4c3ff8ac289a7bff.pdf"}}},{"node":{"id":"Magazine-issue_61b787451e0c2a0a6f33729d","title":"Fall 2005","pdf":{"publicURL":"/static/8aeba504255eb4d976fb9e537c00a299/94b5d64a810e8dce84432849e664217b.pdf"}}},{"node":{"id":"Magazine-issue_6984230ea1914505ee622e4a","title":"Spring 2005","pdf":{"publicURL":"/static/da6100a69acbbe438d23dc9403a7c3ca/855c34211ec59d6d51bd9d239c02be26.pdf"}}},{"node":{"id":"Magazine-issue_61b7877d1e0c2a0a6f3372a2","title":"Fall 2004","pdf":{"publicURL":"/static/b1f7c6a94964aaea6a9c3f798b27586b/dd77cd851c40dc76048c4e32442ca8ba.pdf"}}},{"node":{"id":"Magazine-issue_6984228ea1914505ee622e45","title":"Spring 2004","pdf":{"publicURL":"/static/a99fae12e2f062f751b89341c2f07fba/b39fbbb7eb28dd832fdd285fdaa71c35.pdf"}}},{"node":{"id":"Magazine-issue_69841e8ca1914505ee622e31","title":"Fall 2003","pdf":{"publicURL":"/static/e3082c279b05418209a123fd92fd3c77/a01b72aa38ece43d8bc17de99747673b.pdf"}}},{"node":{"id":"Magazine-issue_69841fd9a1914505ee622e36","title":"Spring 2003","pdf":{"publicURL":"/static/1d8ba0a08350f8678b642bb0ea508aff/e66d14943f0106fea11d11e57d9f4602.pdf"}}},{"node":{"id":"Magazine-issue_69842157a1914505ee622e3b","title":"Fall 2002","pdf":{"publicURL":"/static/47254b4ce38c56c3a1382308412dfd0e/7d87926a4e8aaeb662b695d77e18c882.pdf"}}},{"node":{"id":"Magazine-issue_61b787bc1e0c2a0a6f3372a7","title":"Spring 2002","pdf":{"publicURL":"/static/8709e6fd67151735f2f8dc28df3920f5/b117b2fa8f50b854b04b8f83f994e2aa.pdf"}}},{"node":{"id":"Magazine-issue_6030b07d7782326a48b058c5","title":"Fall 2001","pdf":{"publicURL":"/static/8782e3e443c331d310def3ef5482b484/c7955a146d14a1bb316342ae3d93ecb0.pdf"}}}]}},"pageContext":{"id":"Article_6a1b55f9a1914505ee622e86","categoryList":["6030b07d7782326a48b058be"]}},
    "staticQueryHashes": ["4055801962","4141362474"]}