Michael Hochberg
@mkhochb.bsky.social
530 followers 210 following 100 posts
University of Montpellier - Santa Fe Institute. Working these days on intelligence in biological and artificial systems. Website https://mehochberg.wixsite.com/blog/academia
Posts Media Videos Starter Packs
Pinned
mkhochb.bsky.social
What future does AI hold? With @paulbrainey.bsky.social, we hypothesize that human-AI coevolution is leading to the emergence of a new evolutionary unit of selection. Science fiction or looming fact? Read our paper and judge for yourself. www.pnas.org/doi/10.1073/...
Reposted by Michael Hochberg
khadershabaana.bsky.social
UChicago Microbiology is searching for tenured or tenure-track faculty working in host-pathogens interactions, viral and bacterial pathogenesis, and emerging infectious diseases. Come join our vibrant Department! microbiology.uchicago.edu
Apply here apply.interfolio.com/174404
Apply - Interfolio {{$ctrl.$state.data.pageTitle}} - Apply - Interfolio
apply.interfolio.com
mkhochb.bsky.social
not if you sum it !
jorgeapenas.bsky.social
Antiantifa = fa. It's very simple math.
Reposted by Michael Hochberg
royalsocietypublishing.org
A new study published in #RSOS has uncovered a potential driver behind the aggressive spread of #glioblastoma, one of the deadliest forms of #BrainCancer. Read more: royalsocietypublishing.org/doi/10.1098/...
Reposted by Michael Hochberg
ricardsole.bsky.social
Will humans and AI coevolve in the future? As interdependencies grow, the two could come to function not merely as interacting agents, but as an integrated evolutionary individual. Check this paper in @pnas.org by @mkhochb.bsky.social & P Rainey @sfiscience.bsky.social www.pnas.org/doi/10.1073/...
mkhochb.bsky.social
What future does AI hold? With @paulbrainey.bsky.social, we hypothesize that human-AI coevolution is leading to the emergence of a new evolutionary unit of selection. Science fiction or looming fact? Read our paper and judge for yourself. www.pnas.org/doi/10.1073/...
mkhochb.bsky.social
Very good point. Individual intelligence (i.e., Flynn effect) may or may not beget collective intelligence. Suspect this depends on context, environment, and scale.
mkhochb.bsky.social
What makes intelligence unique? Not surprisingly, no single, simple answer! Progress is however possible on the question if we view intelligence in an information framework. www.sciencedirect.com/science/arti...
mkhochb.bsky.social
Despite the title, I found these interviews informative, particularly those with Eric Schmidt. Main take-aways: AI moving fast, increasing dependence, semi-predictable future. www.youtube.com/watch?v=79-b...
Ex-OpenAI Scientist WARNS: "You Have No Idea What's Coming"
YouTube video by AI Upload
www.youtube.com
mkhochb.bsky.social
yeast cool - and the shirt's probably full of it after a wear to two
wcratcliff.bsky.social
Making a new lab shirt. The yeast settling selection reference is pretty niche, but I don’t mind explaining it to people.

Also, the cells will glow in the dark!
mkhochb.bsky.social
Incredible, but AI is your enemy. Unless I actually see you penciling, I can't be sure that you are the author of this portrait. Come 2030 or thereabouts, original work and AI creation will be indistinguishalbe (clearly I'm not an AI just yet).
mkhochb.bsky.social
Welcome !!!
robjohnnoble.bsky.social
My three months in France have begun! Living in the wonderful, scrubby countryside to the north of Montpellier. A paradise for runners
La garrigue
mkhochb.bsky.social
20/ My hope is that this study will generate interest in intelligence. Can intelligence help understand system organization, behavior and evolution? Will intelligence help in understanding human collective behavior and society? Perhaps most importantly, what is the future of our coevolution with AI?
mkhochb.bsky.social
19/ I have made a preliminary attempt at listing some abilities and the types of systems in which they can be found. There is clearly a LOT to be discussed here, but I believe this is in the right direction.
mkhochb.bsky.social
18/ Understanding the under-the-hood abilities and their interrelations will be central to understanding intelligence, quantifying it, and at some point, developing a theory of intelligence. This is the next step!
mkhochb.bsky.social
17/ The indicators proposed are based on bulk information. They do not say ANYTHING about the nature of information, nor how it was acquired. Thus, my framework is not explicit about what is “under-the-hood”, i.e., the nuts and bolts of intelligence.
mkhochb.bsky.social
16/ Thus, intelligence needs to contextualized, that is, relative to something: the agent’s own ability, the complexity of the task, or to a benchmark (a standardized estimate of complexity). Complex contexts can be equated with goal “difficulty”.
mkhochb.bsky.social
15/ At the other extreme, an agent may have a deep understanding and yet fail to achieve a complex goal, due to, for example, a stupid error.
mkhochb.bsky.social
14/ Intelligence is indicated by how much of the goal-useful information is actually achieved by the agent. Thus, it is possible that the agent has no understanding of a goal and succeeds due to a lucky guess. No information is gained.
mkhochb.bsky.social
13/ Solving and planning “construct” – that is, they amass and organize goal-useful information towards resolution of a goal. More steps, levels and scales, and information-rich sub-goals generally mean greater complexity.