Those of you who are not going to prison for hurting people in a more significant sense than through the oppression surveillance creates, are getting house arrest with ankle bracelet monitors so that you might finally feel viscerally the consequences of your civil liberty violating policies.

I know at least a subset of you and even if you believe yourself to be 'part of the team' a condition of my collaboration with others is that you also are held to account. You did not have my permission to surveil.
Reining in massive exploitation, corruption + civil liberties violations should not take a backseat to "optics". Nor should science innovation or other advances that can contribute to human + whole earth flourishing.

No, I won’t protect or delete my tweets to hide your history of corruption.

You have to live with the your SS behavior—Hacking + tracking, slander, financial shenanigans, gaslighting, attempted coercion + blackmail, using your authority to make fake medical diagnoses, sexual harassment, non-consensual image sharing, entrapment, intimidation, and intellectual theft to get that start-up a’going. And very crappy leadership. For a sense of the institutional, societal + personal consequences of hacking + tracking, watch Florian Henckel von Donnersmarck's The Lives of Others.

Two lovely 🌈 rainbow + one brown trout caught by my brother this past weekend.



The waters chased him as he fled,
Poem 1749 By Emily Dickinson

ASSEMBLY, MICRO-MACRO MAPS + EVOLUTIONARY DYNAMICS @leecronin, @Sara_Imari + @sfiscience colleagues created a fire-storm w a paper @Nature last week that despite limitations makes good progress formalizing the role of compositionality in evolution. 1/n



In doing so they also provided a new complexity measure (note ‘complexity’ is used in multiple ways within the paper itself) that can serve as a signature of an adaptive process (useful for identifying life elsewhere) in so far as it quantifies how algorithmically unlikely (by chance) a given object is + hence how much work is required to produce it. This is nice because it lays more ground work for connecting construction / computation processes in biology to thermodynamics.

The paper has been described as controversial—a better description is "novel but not actually that controversial with a useful framework hidden in a bizarre package". Nearly every framing sentence (as well as many others) of this paper gave me anxiety, not just by virtue of being hard to parse but also because the sentences betray subtle confusions + intellectual inconsistencies—some of which perhaps belong to the authors but many others of which simply reflect the messy frontier. That said, this criticism doesn’t in any sense sink the paper because its true contributions do not depend particularly on its framing. In addition, in a kind of masochistic way I find the authors’ framing informative--its ‘wut’-ness brings to the fore conceptual issues at the intersection of biology and physics that badly need attention.

Below is at take on the paper’s findings + contributions, quotes from the paper for further thought, and flagged issues that need community attention. Before embarking, a final remark is that although this paper—even at this relatively early stage of the work—could be more elegantly and clearly written, I am glad it was published in Nature as these kinds of ideas are important + need more mainstream attention. In contrast to perhaps the conventional view that a journal like @Nature should be publishing the definitive work on some question, I think a better or additional use of such journals is to publish highly generative, thought-provoking or door opening papers, regardless of whether the results or frameworks are likely to stand the test of time.nature.com/articles/s4158…

2/n SUMMARY: the paper (in a mixture of the authors’ language + mine) provides a formal approach for breaking down aggregates into elementary constituents (defined as requiring no assembly—which is a bit of a problem—see below), then recombining these elements in principled ways using assembly rules for when they can be joined to identify the shortest path—least number of algorithmic steps— required to reconstitute/compute the original macroscopic target. The authors call this minimal path the assembly index.

This process also yields the space of possible, counterfactual objects that could be realized from a given elementary set + rule space, but perhaps have not yet been discovered by evolution. The authors then drop the ‘all assembly rules are equally likely (technically drawn uniformly, I think) assumption’, which in implicitly introducing constraints, reduces the space of macroscopic objects that can in principle be computed/assembled, making the approach simultaneously more tractable and realistic.

Dropping this assumption allows introduction + comparison of objects generated by directed vs undirected processes + connects assembly index to selection strength in so far as a high assembly index is marked by high complexity (many steps) and low exploration (low variation due to strong selection acting against drift?). The quantity, ‘assembly’ is a function of the number of copies of an object and its assembly index (recall, that the assembly index is the minimal number of recursive steps necessary to build a macroscopic object, which the authors also call its size).

The authors argue that assembly captures the total amount of selection (or work?) required to produce an ensemble of observed objects and hence how likely an adaptive process is to be operating as well as quantifying its strength. An important note is that the initial rules (or those initially identified by the ‘breaking’ or decomposing process) and elements do not fully determine the space of possible objects—if mass action kinetics are allowed then (as I understand it from the paper) due to a natural symmetry breaking in the construction of contingent assembly paths, a network of growing branches can potentially emerge, changing through novel interactions the space of future objects.

It is here (I think) in the process that ‘biology’ could be said to emerge from physics to take on a life of its own, so to speak, not breaking the laws of physics but making a genuine causal contribution to them + marking an important transition in the universe’s dynamics, a la Wheeler (my take). Mentioned at the end of the paper are also some important points about the role that interacting timescales play in the discovery of complex objects. Also mentioned towards the end is the idea that an arrow of time arises through assembly, with a tick occurring each time an object is made. For more discussion related to the above summary, see also Carl Bergstrom’s excellent explainer thread about the paper on Bluesky.

Best be stopping' the hacking + tracking. We're back on course for integrity. Don't undermine progress with an attempt at censorship or optics. Creativity + quality are allergic to both.

In dark mode now. I will reappear as necessary until this situation is stabilized.

Cheers SS.

https://twitter.com/RealEmirHan/status/1703115467215302786

Misrepresentation + grandiosity would send the wrong message + be disrespectful to the institute, its employees, scientists, + women. Except in the cases of the Hitlers and Stalins and their disciples, it’s of course fine to emphasize the genuinely good. In my experience, owning mistakes + working constructively to repair damage sets an invaluable positive example.

Much love, Cormac.

Looking south from Muley Point at your stomping grounds, where "all preference is made whimsical and a man and a rock become endowed with unguessed kinships." Blood Meridian, 25th Anniversary Edition, page 259.

COMPLEXITY BEGETS COMPLEXITY: Improvements in measurement tech + exp design reveal complexity across scales, challenging the idea it increases in more than a Russian doll sense from cells to societies. Below—an old thread + new insect/bees results. 1/N

https://twitter.com/C4COMPUTATION/status/1217254558209961985?s=20

Cultural learning (diffusion of neutral behavioral variants) was thought limited to large brained animals. @alicedbridges @LChittka show @PLOSBiology bees socially learn to solve puzzle boxes+if exposed to mult equally good solutions nonetheless come as a group to prefer one. 2/N

Happy Birthday, Claude Shannon.

One of Shannon's conceptual feats was to connect information content to uncertainty via entropy. Uncertainty reduction has become one of the most powerful concepts in biology showing up as a foundational idea in cognitive science + evolutionary theory via predictive coding, theories of individuality + collective computation + fitness. 72 years later other components of “information”—semantics + function—await formalization.

Weaver on three "communications problems," from their 1949 monograph: raley.english.ucsb.edu/wp-content/Eng…

Among the most basic science questions in evolutionary biology are: how do levels of organization arise, does complexity increase in an interesting way as we move up scales, + do the levels “talk” to each other: does what happens at the microscale or individual level matter 1/8 to interactions at the macro scale? @MarioMuscarella +@Jp_odwyer creatively address some of these issue in . They show by building a coarse-grained model of a species’ cellular metabolism several counterintuitive results, one of which is 2/8 tinyurl.com/1ll2hgzq

Complexity all the way down: It's becoming clear that the view complexity increases as we move up life's organizational hierarchy is wrong (thread)

Dendrite compartments allow single neurons to compute XOR @mattlark @QuantaMagazine
tinyurl.com/tyreke5
tinyurl.com/sfh3kn4 @mattlark @QuantaMagazine Complexity all the way down: Bacteria + other prokaryotes have complex internal structure w membrane bound organelles that can sequester genetic material, allow nutrient digestion + synthesize molecules. Read more in this accessible @QuantaMagazine essay tinyurl.com/sxccg46