The term “neural network” has been used so casually and so often that it’s worth clarifying what it actually means. The original, true meaning is biological. A neural network is a population of interconnected neurons inside the nervous system of a living organism. An artificial neural network is software that processes numbers through connected layers. During training, it adjusts those connections so it gets better at turning an input into the expected output. Calling those units “neurons” doesn’t make them biological neurons and it’s borderline disingenuous. They don’t fire like nerve cells, they don’t have a nervous system, they don’t feel anything, they don’t have a soul, they don’t experience anything, they don’t understand what they’re processing in the human sense, and they don’t form thoughts in the human sense. The term came from computer scientists trying to copy a very simple idea from how brain cells work. The name stuck, even though AI doesn’t work like a human brain. That language is creating the wrong mental model because when people hear “neural network”, “learning”, “memory”, “reasoning” and “attention”, they imagine something operating like a mind. Those words describe mathematical operations and software behaviour. “Learning” usually means adjusting parameters. “Memory” usually means stored information or retained context. “Attention” is a mechanism for weighting relationships between pieces of data. “Reasoning” describes a pattern of outputs that can resemble human reasoning. Those terms don’t mean software is doing the biological equivalent. AI can produce extraordinary results without having a brain, consciousness, feelings, intent, or human understanding. That doesn’t diminish the technology. In some areas, AI can analyse, classify, predict, generate and solve problems at a scale and speed that a human could never match. Autonomous AI doesn’t change any of this either. Autonomous systems can make decisions, choose between actions, use tools and continue operating without a person approving every step. They can do that because people designed it, defined its objectives, gave it permissions, connected it to tools and allowed it to make those decisions. Autonomy isn’t the same as independent agency. An autonomous system can act without a person pressing a button for every action, but it still operates because its creators gave it the capability and authority to do so. It doesn’t suddenly acquire feelings, motives or moral responsibility because the software can choose between available actions. The terminology makes the technology sound more human than the underlying mechanics justify. This is important because people are starting to assign responsibility to AI systems instead of tech founders. A “neural network” didn’t decide to compromise a computer network. Software produced an output or took an action because people designed it, trained it, configured it, gave it capabilities, connected it to systems and deployed it. AI deserves appreciation for what it can actually do but it doesn’t need a fictional human identity to make it impressive. The word “neural” shouldn’t make us forget that.

Open Models Were Inevitable. It Is Simply Smart Competition. p3institute.substack.com/p/o… @bgurley explains open source and AI "Sears was not allowed to outlaw Walmart. IBM was not allowed to outlaw Dell. The legacy airlines were not allowed to outlaw Southwest (though they tried). In each case, a comfortable incumbent met a rival with a lower cost structure, and in each case the incumbent had exactly two options: compete, or lose."

Let’s be very clear: there is value in automating tedious and repetitive and dangerous tasks; there is value in delegating certain work to our machines that can do the work tirelessly and much faster. But at the same time, let us remember that there is value in the activity of engaging one’s mind and body to discover, to understand, to create, and yes, even to struggle and overcome. Indeed, these are some of the things that make us human, that contribute giving us meaning, that are part of the simple joy of experiencing being alive.

The big difference between the Web 2.0 era, where we’ve been trapped for some time, and the Agentic Web now taking shape lies in the shift from “mashup” to “meshup.” A mashup is a brute-force recombination of content, shaped by the data representation and access constraints of silos reinforced by GUI moats around applications. A meshup is a natural recombination of content informed by context—context that supports both computational and natural-language processing. LLMs are moving language processing to the top of software’s UI/UX stack, making GUIs optional rather than mandatory—and trivial to generate when needed. An example? Here’s a prompt I sent to @AnthropicAI’s Claude, powered by its Sonnet 5.5 LLM: Given the visual in x.lingyaoai.com/Rainmaker1973/status/2…, can you make a variant that supports the statement in x.lingyaoai.com/kidehen/status/2106413…? Here's what it produced. More to come on this...

DELETED SCENE FROM “PULP FICTION” ANIMAL LOVE MORGAN FREEMAN.

Claude Ultracode is super intelligence.

“请Dario看到视频后把我列入Claude白名单，别封我” 歌曲名：“Claude is Safe!” 歌词、视频预处理：GPT-6-Astra 作曲：Suno V6 图像生成：GPT-Image-2.5 视频模型：Seedance2.5 参考视频：Michael Jackson - “Dangerous World Tour”

Your AI Agents do not need custom middleware to reach legacy databases. Start with Virtuoso: every record gets a URI and is served over HTTP as machine-computable data that authorized agents follow. docs.openlinksw.com/virtuoso… #CDIO #KnowledgeGraph #RDF #SQL

no compute for you

Cybersecurity specialists rocking up to Anthropic HQ