To ChatGPT: How does the Neuralink Blindsight brain implant in visual cortex compare with the ReVision Implant brain implant in visual cortex? What would be the best alternative for restoring a form of vision to the blind? https://chatgpt.com/share/68546124-0dd0-8004-9d3e-ff98c122e468 The **Neuralink Blindsight** and **ReVision Implant** both aim to restore vision through **visual cortex stimulation**, but they differ markedly in technology, goals, transparency, and scientific grounding. Here's a **side-by-side comparison** based on current available information: --- ### 🔬 **Core Concept** | Feature | **Neuralink Blindsight** | **ReVision Implant** | | -------------------- | ------------------------------------------------------------------------------- | --------------------------------------------------------------- | | **Stimulation Site** | Primary visual cortex (V1) | Also primary visual cortex (V1) | | **Purpose** | Restore some visual perception (e.g., phosphenes) to the blind via camera input | Same: create usable visual percepts from electrical stimulation | | **Target Users** | Primarily blind individuals (initially late blind) | Similar, focused on individuals with no usable vision | --- ### 🧠 **Implant Technology** | Feature | **Neuralink** | **ReVision** | | ----------------------- | -------------------------------------------------------------------- | ------------------------------------------------------------- | | **Electrodes** | \~1,024–3,072 ultra-thin polymer threads | \~65–96 electrodes per array (based on Utah-like arrays) | | **Placement** | Inserted via robotic microsurgery to reach layer 5 of V1 | Implanted via open craniotomy, electrodes penetrate cortex | | **Depth of Electrodes** | Flexible threads inserted several mm deep | Utah array-style: \~1–1.5 mm stiff electrodes | | **Form Factor** | Coin-sized disk (“Link”) flush with skull, wires fan out into cortex | Multiple stiff array modules (e.g. 4 arrays of 96 electrodes) | --- ### ⚙️ **Functionality & Limitations** | Feature | **Neuralink** | **ReVision** | | ---------------------- | ------------------------------------------------------------------ | ---------------------------------------------------------------- | | **Vision Output Type** | Sparse phosphene clouds (e.g. stars, dots, lines) | Similar phosphene-based vision | | **Resolution** | Limited due to number of active electrodes & safety (seizure risk) | Also limited, but uses higher current for fewer phosphenes | | **Camera Interface** | External camera (possibly in glasses), processed by onboard chip | Similar camera interface feeding processed images to cortex | | **Neural Coding** | Proprietary encoding via AI processing; few details disclosed | Explicit use of stimulation maps based on phosphene localization | --- ### 🧪 **Scientific Grounding & Testing** | Feature | **Neuralink** | **ReVision Implant** | | ---------------------- | --------------------------------------------------------------------------------- | ------------------------------------------------------------- | | **Peer-reviewed Data** | **None yet released** on visual cortex use | **Published animal studies** and human feasibility trials | | **Human Testing** | As of mid-2025: 1st human with Link in motor cortex, **not yet in visual cortex** | Feasibility trials in blind human subjects (e.g., at UCLA) | | **Transparency** | Minimal details; more PR than science | Academic collaboration (e.g., with Richard Andersen, Caltech) | --- ### ⚠️ **Technical & Safety Issues** | Concern | **Neuralink** | **ReVision Implant** | | ----------------- | -------------------------------------------------------- | ------------------------------------------------------------- | | **Seizure Risk** | High if too many electrodes activated simultaneously | Also high, hence sparse activation | | **Tissue Damage** | Low per electrode (soft threads), but many insertions | Higher due to stiff electrode shafts penetrating cortex | | **Longevity** | Unknown, possibly limited by gliosis and thread movement | Known limitations from Utah arrays: inflammation, scar tissue | --- ### 💡 **Unique Aspects** | Feature | **Neuralink** | **ReVision Implant** | | ------------------- | ------------------------------------------------------- | ------------------------------------------------- | | **BCI Ambition** | Part of broader BCI vision (telepathy, control, memory) | Focused on visual restoration only | | **Surgical Method** | Proprietary robotic system inserting thin threads | Manual placement of electrode arrays | | **Innovation** | Flexibility of threads + robotic precision | Earlier and more scientifically grounded attempts | --- ### 🧾 **Bottom Line** | | **Neuralink Blindsight** | **ReVision Implant** | | ------------------ | --------------------------------------------------------- | ------------------------------------------------------- | | **Current Status** | Vaporware for vision restoration as of 2025; demo pending | Limited but **scientifically demonstrated** feasibility | | **Strengths** | Cutting-edge electrode tech; scalable | Academic transparency; first-in-human data available | | **Weaknesses** | No real data; possible overpromising | Stiff electrodes cause long-term damage; low resolution | --- ### 🧠 Summary Judgment * **Neuralink Blindsight** is **theoretically more scalable** and less damaging per electrode, but it is **still unproven for vision** and lacks open validation. * **ReVision Implant** is **more grounded in published research**, has **human trial experience**, but faces physical limitations due to older electrode technology. If your goal is real, current restoration of vision—even limited—**ReVision is ahead**. If you’re betting on the **future of scalable cortical vision BCIs**, **Neuralink’s platform** may hold more long-term potential, but it's **years behind in the visual domain**. Would you like an **illustrated comparison infographic** showing these differences visually? %%% Me to ChatGPT: What would be the best alternative for restoring a form of vision to the blind? The **best current alternative** for restoring a **form of vision** to the blind—particularly for those who are totally blind—is **non-invasive visual-to-auditory sensory substitution**, especially the system called: --- ### 🎧 **The vOICe** – Visual-to-Auditory Sensory Substitution System --- #### ✅ **Why it’s the best available alternative:** | Feature | Explanation | | ------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Non-invasive** | No surgery, no implants, no infection or glial scarring risk. Runs on a smartphone with camera glasses and headphones. | | **Information-rich** | Translates full camera video into continuous soundscapes encoding visual shape, position, size, perspective, and even texture. | | **High resolution** | Unlike brain implants limited to dozens–thousands of stimulation points, The vOICe leverages the full resolution of a camera and human auditory system. | | **Real-time** | Offers live, uninterrupted visual information with no delay—no dependence on AI "scene understanding" hallucinations. | | **Neuroplasticity-supported** | Trains the visual cortex through auditory input, effectively “recruiting” it for vision without needing physical implants. fMRI studies show activation of V1 in blind users. | | **Works for the congenitally blind** | Unlike most visual prosthetics, The vOICe can work even if the user has never seen before. | | **Open access** | It is scientifically documented and available as a free app for Android, developed by Dr. Peter Meijer. | --- ### 🧠 Compared to Other Options | Technology | Invasiveness | Resolution | Availability | Suitable for Congenitally Blind? | Real-world Utility | | -------------------------------------------- | ---------------------- | --------------------------------- | ------------------------------ | ------------------------------------- | ------------------------------------------------------------- | | **The vOICe** | None | High (camera resolution) | Yes, free or low-cost | Yes ✅ | Proven with long-term users | | **Brain Implants (Neuralink, ReVision)** | Highly invasive | Low (phosphenes only) | Experimental | No ❌ (limited by cortical plasticity) | Extremely limited so far | | **Retinal Implants (Argus II, PRIMA)** | Invasive (eye surgery) | Very low | Discontinued or limited trials | No ❌ | Largely abandoned | | **AI Smart Glasses (e.g., Envision, OrCam)** | Non-invasive | Medium (scene/label reading) | Commercially available | Yes ✅ | Good for OCR, face, scene labels—but **not for "raw" vision** | | **Echolocation / Cane** | None | Low (but good for depth/distance) | Universally available | Yes ✅ | Great for navigation but not object recognition | --- ### 🚫 Why Brain Implants Fall Short * **Very low resolution**: Visual cortex implants produce only sparse phosphene dots, not detailed vision. * **Limited simultaneous stimulation**: To avoid seizures, only a few electrodes can be active at once. * **High surgical risk**: Gliosis, infection, electrode failure. * **No natural encoding**: The brain must interpret abstract dot patterns, not visual scenes. --- ### 💡 Ideal Future Vision Support In the foreseeable future, the **optimal approach** is likely to be **hybrid**: > **Smart Glasses with The vOICe + AI overlays** > → Raw sensory substitution with interpretable AI assistance (labels, OCR, navigation). This allows the user to **maintain control**, develop genuine perception skills, and get AI support **only when they choose to ask for it**—avoiding overdependence and mental “dumbing down.” --- Would you like an image or infographic to communicate this visually? %%% Me to ChatGPT: Yes. https://chatgpt.com/s/m_685475411134819190094e381e9d3ca9