现实与科幻小说之间的界线日益模糊，特别是在「脑电波读取」与脑机介面（BCI）领域。旧金山创业家暨投资人 Saturnin Pugnet 在其名为《Phase Lock》的论文中指出，脑机介面技术正以惊人的速度发展，未来有望升级人类的「软体」并赋予人类认知超能力。在人工智慧加速发展的时代，Pugnet 认为既然无法减缓 AI 的进化，就必须透过神经技术为人类「加速」，将生物智慧与数位智慧结合以提升人类的大脑频宽，这种观点也反映了马斯克创立 Neuralink 背后的矽谷共识。（关键数字： one day）
受人工智慧进步的推动，神经技术领域的创投融资从 2024 年的 23 亿美元翻倍至去年的 48 亿美元。Pugnet 主张，神经技术在未来十年可以受益于过去推动大型语言模型（LLM）发展的「规模化法则」（scaling laws）。透过投入庞大的运算能力与脑部扫描数据，AI 模型将能够识别并复制大脑活动的模式，不仅能让研究人员彻底理解大脑并增加新功能，还能借由训练 AI 学习脑机介面收集的数据，实现比现有控制方法更可靠的「整合式对齐」（alignment by integration），进而开发出更符合人类价值的安全 AI。
目前，侵入式与非侵入式脑机介面技术皆在医疗领域取得显著进展，例如帮助瘫痪患者以每分钟 56 个字的速度进行语言与游标控制，以及英国新创公司 Cerca Magnetics 获得政府资助在医院安装非侵入式脑部扫描设备以治疗癫痫。尽管如此，部分神经技术专家仍对规模化法则的效果持保留态度，认为大脑信号会随情绪或生理状态而剧烈变化，电脑要读取抽象思维甚至实现双向「读写」仍充满挑战，但创投的资金投入即便未能完全实现此愿景，也能为医学研究留下宝贵的数据资源。
The boundary between reality and science fiction is increasingly blurring, particularly in the realm of "brain reading" and brain-computer interfaces (BCIs). In his essay titled "Phase Lock," San Francisco entrepreneur and investor Saturnin Pugnet explains that BCI technology is advancing rapidly and could one day upgrade human "software" to grant cognitive superpowers. In an era of accelerating artificial intelligence, Pugnet argues that since we cannot slow down AI, we must speed up humans through neurotechnology by merging biological and digital intelligence to increase human bandwidth—a view that echoes the Silicon Valley consensus behind Elon Musk's launch of Neuralink.
Driven by advances in AI, venture funding for neurotechnology more than doubled from $2.3bn in 2024 to $4.8bn last year. Pugnet contends that neurotechnology over the next decade can benefit from the same scaling laws that propelled large language models (LLMs). By applying massive computing power to brain-scanning data, AI models could recognize and replicate patterns in brain activity, enabling researchers to fully understand the brain, add new capabilities, and achieve "alignment by integration"—training AI on BCI-amassed data to develop safer models that reflect human values more reliably than current control-based methods.
To date, both invasive and non-invasive BCI technologies are making significant strides in medicine, such as enabling a paralyzed patient to communicate at 56 words per minute and supporting UK start-up Cerca Magnetics with government funding to deploy non-invasive brain scanners for epilepsy treatment. Nevertheless, some neurotechnologists remain skeptical of the scaling approach, noting that brain signals vary wildly depending on a person's emotional or physical state, making it a leap of faith to believe computers can read abstract thoughts or achieve bidirectional "reading and writing"; still, venture capital efforts could yield precious data resources for medical research even if they fall short.