The unbearable slowness of being: Why do we live at 10 bits/s?

Access and Context

• Multiple commenters note the journal paywall; several link to the free arXiv preprint.
• Some urge others to read the full paper before judging, while others say the press coverage overstates what is actually done.

Headline Claim: 10 bits/s Conscious Throughput

• Central claim discussed: conscious “behavioral throughput” ≈ 10 bits/s vs ≈10⁹ bits/s sensory input.
• Many find 10 bits/s intuitively far too low, citing speech, reading, typing, gaming, or sports.
• Defenders stress this is about a narrow, compressed, high‑level cognitive bottleneck, not raw sensing or reflexes.

Information Theory and Language Rates

• Repeated reference to Shannon’s estimate that English has ~1 bit/character, leading to ~10 bps at ~120 WPM typing.
• Others point out extreme text compression (e.g., Wikipedia at <1 bit/char) and argue conscious semantic content is even sparser.
• Critics respond that this “compression framing” can always be tuned to hit 10 bps, making the number feel arbitrary or clickbaity.

Methodology and Examples

• 20 Questions: many argue it is a poor basis for a cognitive rate, being social, cooperative, task‑dependent, and highly compressed.
• Rubik’s cube, StarCraft, reaction tests: critics say mapping actions or APM to “bits” is reductive and misinterprets skill, pattern recognition, and motor learning.
• Some note complex problems can be stated in few bits (e.g., math integrals) but require huge computation, so “bits out” ≠ “processing done.”

Serial vs Parallel, Conscious vs Unconscious

• Paper’s use of dual‑task “psychological refractory period” to argue central serial processing is challenged; others cite multitasking and movement research suggesting more flexibility.
• Widespread agreement that most processing is unconscious and massively parallel; consciousness may be a narrow, serial “framebuffer” or attention stream.

Analogy Limits and Use of “Bits”

• Several object to treating humans as digital systems; argue continuous, embodied, biochemical processes don’t map cleanly to bits.
• Others counter that information‑theoretic bits (entropy) are substrate‑agnostic and legitimate for high‑level capacity estimates, if used carefully.
• Reductionism is defended as necessary but also blamed when it yields seemingly nonsensical numbers like “10 bits/s.”

Neural Interfaces and Practical Implications

• The paper’s jab that a Neuralink‑style BCI may be no better than a telephone (if cognition is 10 bps) is widely quoted; some find it funny and thought‑provoking.
• Others argue that shortening feedback loops and delegating to AI agents might still greatly boost productivity even with a slow human bottleneck.

Paper Quality and Media Hype

• Some praise it as a stimulating perspective that synthesizes data and poses useful questions about inner vs outer brain and routing/attention.
• Others call it “blog‑post level,” over‑reductive, and poorly grounded, criticizing back‑of‑envelope math and lack of new measurements.
• Media coverage using terms like “measure” and “quantifies” is seen as overstating what is, at best, a speculative, order‑of‑magnitude framing.