Summary of "Наш интеллект УМИРАЕТ. Как ИИ разрушает сознание? | Нейробиолог Алипов, Михаил Никитин"

Cognitive effects of tools and AI

Externalized memory and cognition
- Tools that externalize memory — writing, search engines, calculators, and AI — reduce selection pressure for large internal memory capacity and change which cognitive efforts are valuable.
- These technologies shift effort from rote memory toward skills like retrieval, synthesis, evaluation and tool‑use.
Impact of AI
- Large language models and generative media amplify the capabilities of already higher‑skilled users, while also enabling shortcuts, passive consumption, and potential loss of practiced skills.
- AI can improve problem solving and bring some learners closer to average performance; a 2024 systematic review found overall benefits of AI‑assisted problem solving, though some studies had biased samples.
- AI offers individualized tutoring at scale, but widespread use also creates cascading evaluation problems (e.g., applicants and screeners both using AI) and new trust problems (deepfakes, voice/face synthesis).
Potential social risks
- Scenarios of extreme change are discussed (e.g., neural interfaces coupled with ubiquitous LLMs), but AI may also level up students and provide individualized teaching.
- Small psychiatric vulnerabilities combined with persuasive AI could lead to conspiracy beliefs, induced psychosis, or reverential attitudes toward AI.

Evolution, intelligence and genetics

Historical trends and the Flynn effect
- The Flynn effect describes large IQ gains across many countries during the 20th century, attributed to factors such as improved nutrition and schooling (notable gains in Japan and China).
Genetic architecture of intelligence
- Intelligence is highly polygenic. Rare variants that appear to increase IQ are often rare because they carry trade‑offs or non‑additive interactions with other variants.
- High IQ scores are not straightforwardly comparable across domains; IQ tests were designed to detect low performance and have limits at the extremes.
Psychiatric correlations and trade‑offs
- High intelligence can correlate with increased neuroticism or anxiety. Some psychiatric disorders reduce measured cognitive performance.
- Natural selection tends to increase beneficial common variants; rare “IQ‑boosting” variants often indicate pleiotropic costs (e.g., epilepsy, neuroticism), making simple genetic “super‑intelligence” routes unlikely.
Cultural and evolutionary feedbacks
- Technologies and cultural changes that reduce the difficulty of survival or success for many people can reduce the selective advantage of larger brains (cultural‑drive hypothesis).
- Biological trends (e.g., brain‑size changes) operate over many generations and are therefore slow.

Autism and neurodiversity

Reasons for increased autism diagnoses
- Improved recognition and broader diagnostic criteria have raised detection rates.
- Demographic effects such as increased paternal age lead to a higher burden of de novo mutations in sperm (spermatogonial divisions continue with age), increasing genetic risk.
- Modern, information‑dense and urban environments can exacerbate phenotypes that were less problematic in small‑scale or preindustrial settings.
Spectrum and strengths
- Autism is a spectrum: many autistic people have normal or above‑average intelligence.
- Traits such as hyperfocus and intense special interests can enable deep expertise and successful careers, especially in IT and engineering.
- Hypotheses about altered pruning or connectivity may underlie some neurodevelopmental differences.
Practical mitigation note
- Sperm cryopreservation at younger ages is discussed as a mitigation strategy for reducing age‑related de novo mutation risks; this approach has limitations related to storage, cost and other factors.

Social and behavioral impacts of sex‑tech and AI companionship

Changes in social norms
- Sex robots and AI companions may normalize submissive or idealized interaction modes and potentially weaken interpersonal relationship skills through practiced, nonreciprocal interactions.
Psychiatric and social risks
- Persuasive AI combined with individual vulnerabilities could contribute to pathological beliefs or reverence toward AI entities.
- AI also increases opportunities for social engineering: scams, synthetic persona abuse, and misleading content on dating platforms and social media.

Life beyond Earth and biosignatures

Recent claims and historical context
- A 2024 Nature paper reported unusual spectral and chemical signatures from Perseverance rover samples in a Martian crater that might be biosignatures.
- Earlier claims (for example, the 1996 ALH84001 meteorite) were controversial; abiotic processes can produce similar features, so such claims require cautious interpretation.
Confirmation requirements and habitats
- True confirmation of extraterrestrial life would require sample return and laboratory analyses on Earth.
- Subsurface Martian habitats (deep underground) are plausible locations for extant life.
Historical lessons from chemistry
- Early ideas that organic molecules necessarily implied life were overturned by laboratory syntheses (Wöhler, Butlerov); many organic molecules form abiotically.

Blockquote:

“Abiotic processes can mimic many biosignatures, so extraordinary claims require sample return and careful terrestrial analysis.”

Convergent evolution of nervous systems and cultural transmission

Independent origins and convergence
- Complex nervous systems and cognition have evolved multiple times; cephalopods (octopuses) developed complex brains independently and show convergent features such as ventricles and protective capsules.
Cumulative culture
- Cumulative culture requires reliable copying with room for innovation. Humans achieved large‑scale cumulative culture because of dense social networks, long‑distance exchange, and high‑fidelity learning.
- Nonhuman examples of cultural transmission include learned tool use in birds, macaques, and socially learned cooperative hunting in wolves; archaeological evidence (for example, traded cowrie shells) documents long‑distance cultural exchange in prehistory.
Ecological constraints and alternative strategies
- There are ecological and physical limits on evolving traits such as raw strength or speed; cognition and social strategies offer alternative adaptive routes.

Brain size and tempo of change

Long timescales
- Human brain size increased substantially over roughly 2.5 million years. Recent modest reductions in brain size reported in some populations are slow, multigenerational trends.
- Drivers may include cultural niche effects, reduced selection for large brains, and the high energetic cost of brain tissue.
- Noticeable genetic shifts require many generations (tens to hundreds or more); short‑term phenotypic changes often reflect environmental and cultural effects.

Mechanisms and cause‑and‑effect summaries

Reasons for apparent rise in autism diagnoses
1. Improved diagnosis and awareness.
2. Increased paternal age → more de novo mutations in sperm → higher genetic risk.
3. Modern information‑rich, urban environments exacerbating phenotypes.
Mechanisms that could reduce selection for high cognitive investment
1. Widespread inexpensive technologies that obviate cognitive labor (writing → external memory; internet/AI → external problem‑solving).
2. Cultural memes/practices that occupy cultural space without increasing biological fitness.
3. Reduced reproductive differentials favoring high‑IQ individuals.
4. Long‑term cultural simplification lowering pressure to maintain large brains.
Why rare “IQ‑boosting” variants are unlikely easy routes to super‑intelligence
1. Beneficial variants should be enriched by natural selection if common and without cost; rarity often signals trade‑offs.
2. Genomic effects are highly polygenic and non‑additive; combining rare variants can cause deleterious pleiotropy (e.g., epilepsy).
3. IQ tests are limited and domain‑specific; extreme scores may reflect narrow strengths rather than universal ability.
How AI is already affecting education and work
- Helps solve cognitive and analogical problems across ability levels (study quality varies).
- Enables personalized explanation and individualized tutoring at scale.
- Facilitates automation of resume writing, coding and Q&A, producing cascading evaluation issues when both applicants and screeners use AI.
- Increases social trust problems (scams using synthetic voice/face/identity and AI‑written messages on dating platforms).

Researchers, sources and references mentioned

Podcast guests and host
- Mikhail Alexandrovich Nikitin — biologist (Moscow State University)
- Vladimir Ivanovich Alipov — neurobiologist / physician
- Host: Gleb Solomin (podcast)
Papers, studies and historical references
- Flynn effect (James Flynn; literature on 20th‑century IQ gains)
- 2024 systematic review/paper on cognitive abilities and artificial intelligence (unnamed exact citation)
- 2024 Nature paper on unusual chemical/spectral signatures from Perseverance rover samples
- 1996 Science paper on the ALH84001 Martian meteorite (claimed biosignatures; later disputed)
- Literature on paternal‑age effects and de novo mutations related to autism risk
- Historical chemistry: Friedrich Wöhler (urea synthesis), Aleksandr Butlerov (chemical synthesis concepts)
- Additional mentions: Terence Tao, Lex Fridman, Pavel Durov, Mensa, Perseverance rover and sample‑return mission plans

Additional notable natural and biological examples

Convergent evolution in octopuses — independently evolved complex brains.
Magnetotactic bacteria — featured in debates over abiotic versus biotic origins of structures.
High‑performance insect flight muscles (fly flight‑muscle power) as extreme physiological examples.
Archaeological evidence of long‑distance trade, e.g., tropical cowrie shells found in ancient graves in the Russian plain.