What Dolphin Singularity is
and why it exists
Marine cognition research is advancing fast. The last five years have brought AI-powered acoustic analysis, large-scale population studies, and serious scientific attention to questions that were once considered fringe. The problem is that the research is scattered — it lands in journals you don't follow, preprints that don't make the news, lab blogs that disappear, conference proceedings that never get indexed.
Dolphin Singularity is the synthesis layer. A weekly newsletter covering two or three recent developments in cetacean communication, animal cognition, or AI-meets-biology research — written the way I actually think about it, not the way a press release does.
Why cetacean communication, and why now
Dolphins have been studied for decades, but the field has historically been underfunded, methodologically inconsistent, and vulnerable to both overclaiming (John Lilly in the 1960s) and underclaiming (the reflexive skepticism that followed). What's changed recently is the tooling. Machine learning models can now process acoustic data at a scale that was previously impossible. Google's DolphinGemma project is the most visible example, but it's part of a broader shift: bioacoustics is becoming a data science problem, and the field is moving accordingly.
That shift creates a gap. The people who understand the AI don't always understand the biology. The people who understand the biology don't always follow the AI literature. And most of the popular coverage serves neither group — it either oversimplifies the science or sensationalizes results that don't warrant it.
The newsletter exists to occupy the space between "dolphins have names" headlines and primary literature that assumes you have a background in acoustics signal processing.
The editorial angle
This is not a neutral aggregation service. Every issue has a point of view. When a study is being overhyped, I say so. When a finding is genuinely significant and underreported, I explain why. When two results are in tension with each other, I work through what that actually means rather than treating them as separate stories.
I'm also willing to engage with the philosophical dimension. Questions about what cetacean communication research implies for animal consciousness, for cognition more broadly, and for how we should think about non-human minds — these aren't separable from the science. They're part of why the science matters.
What I try not to do: moralize without substance, project human frameworks onto animal behavior without flagging it, or treat accessibility as an excuse for imprecision.
Format
One email per week, roughly 800 words. Usually two or three items — a recent paper or preprint, a development in the AI/bioacoustics space, and something from adjacent fields (comparative cognition, philosophy of mind, conservation policy) when it's relevant.
The free tier gets every issue. After eight weeks of public archive, full back issues become paid-only at $8/month. University departments and graduate researchers can request free institutional access — no strings attached.
The name
"Dolphin singularity" is a half-serious term for the threshold moment when AI-assisted translation of cetacean communication becomes genuinely bidirectional — when we stop just classifying dolphin sounds and start being able to respond in kind. That moment may be closer than it sounds, or it may be further than current hype suggests. Either way, tracking the path toward it is interesting.
The name also captures something about the broader category: the convergence of AI capability and biological complexity. Dolphins are a compelling case study, but the questions the research raises — about non-human intelligence, about what language requires, about how we recognize minds unlike our own — extend well beyond them.
Contact
Corrections, paper recommendations, institutional access requests, and disagreements are all welcome at noah@dolphinsingularity.org. I read everything, though I can't reply to all of it.
What's on this site
Newsletter
The main product. Landing page with pitch, pricing, sample issue, and Substack embed.
Reading List
20 annotated books on cetacean communication, animal cognition, AI/bioacoustics, and philosophy of mind.
Blog
Extended pieces on specific topics — signature whistles, DolphinGemma, the ethics of interspecies communication.
Research
Key papers, projects, and research threads in cetacean communication and bioacoustics AI.
Glossary
Plain-English definitions of terms that come up repeatedly: signature whistles, echolocation, DolphinGemma, and more.
Conservation
The policy and conservation dimension — what the cognitive science implies for how we treat cetaceans.
Editorial positions, stated plainly
- Classification is not translation. Identifying that a sound belongs to a category is different from knowing what it means. A lot of current AI bioacoustics work conflates these, or the coverage does. I try to be explicit about which kind of claim is being made.
- The hype cycle is real and damaging. When results get overstated — and they do, routinely — it erodes credibility for the legitimate research that follows. I'm more interested in accurate calibration than in making the field sound more advanced than it is.
- Animal cognition research has a methodological debt to repay. Decades of poorly controlled studies, anecdotal evidence dressed up as data, and anthropomorphism treated as hypothesis have made serious researchers defensive. That defensiveness sometimes goes too far in the other direction. Both tendencies are worth tracking.
- The philosophical questions are real questions, not decoration. What does it mean for a non-human animal to have a self-concept? What does communication require? These aren't soft add-ons to the science — they determine what counts as evidence and what we're actually trying to find out.
- Accessibility doesn't require imprecision. You can write clearly about complex subjects without dumbing them down. I assume readers who are curious and intelligent, not necessarily credentialed. The newsletter is calibrated accordingly.
Read the first issue free
If the above sounds like the kind of newsletter you'd actually read, subscribe on Substack. Free to start, $8/month if you want the full archive and want to support the work.