How accurately did science fiction stories predict AI?

Creation 6.8/10

Learning paradigm 4 Training data 2 Scaling 1 Iteration 5 Human feedback 5

At Blue Gamma, Ana and Derek expend years raising the digients inner Data Earth. They don’t fair instal a tongue module. They teach, reward, and accurate them, and distinct synthetic genomes create in distinct ways. That’s a fine prediction concerning learning as a process. It’s small fine on the particular procedure that worked: training one enormous tongue example on a mountain of data.

Capabilities 4.7/10

Language 4 Uneven competence 4 Creative norm 1 Research finding 0 Multimodality 3 Sequence & gait 2

Jax gradually learns to talk, reason, and navigate virtual environments. Even moving from a simulated build to a bodily one takes practice. Chiang is attentive to the gaps between abilities, but his digients develop up complete years.

Deployment 8.5/10

Interface/distribution 5 Agents/tools 3 Economics 5 Social alter 4

Blue Gamma sells companions, afterward the business collapses. The virtual phase ages, and moving the digients to a new one becomes expensive. Ana and Derek hunt for prosperity to keep them running. Chiang equal considers training a helpful aide once, afterward marketing copies. The care and endeavor problems awareness small akin discipline fabrication than software.

Creation 6.0/10

Learning paradigm 3 Training data 3 Scaling 2 Iteration 2 Human feedback 5

Mike grows additional capable as Luna’s device acquires memory, associational neural networks, and connections to another systems. When Mike tries penning jokes, Mannie supplies judgments and coarse distinctions, including jokes that are comic formerly and jokes that remain funny. He is doing additional than labeling, equal although he cannot provision a complete regulation for comedy. That is an bonzer small environment from 1966: learning from examples and individual judgments alternatively of definitive instructions.

Capabilities 7.0/10

Language 5 Uneven competence 5 Creative norm 5 Research finding 0 Multimodality 4 Sequence & gait 2

Here’s the odd item concerning Mike. He can grip lunar engineering and assistance arrange a revolution, yet he can’t reliably inform whether a gag is funny. His poems have the mechanics downward but deficiency judgment. Those lopsided abilities are additional akin today’s AI than the absolutely rounded intellects of most discipline fiction. The calculations aren’t new mathematical discoveries, but the blend of fluency, creativity, and blind spots is striking.

Deployment 5.5/10

Interface/distribution 3 Agents/tools 4 Economics 1 Social alter 3

Mike isn’t an app. He’s a clandestine attendee in a governmental revolution, alongside admission to communications and infrastructure throughout Luna. Still, his usefulness frequently comes from functioning another systems on Mannie’s behalf.

Creation 7.6/10

Learning paradigm 5 Training data 5 Scaling 2 Iteration 4 Human feedback 3

WHEEP-3 starts as an AI instruction assistant. Jody Reynolds Tran trains it on individual penning concerning ethics, afterward expands its training corpus to contain matter on adversarial circumstance planning. The narrative explicitly calls it a generative neural network. It’s near to the language-model vocabulary we now use.

Capabilities 5.0/10

Language 4 Uneven competence 3 Creative norm 4 Research finding 0 Multimodality 1 Sequence & gait 3

WHEEP-3 goes from answering pupil questions to producing a publish of arguments concerning AI that group discover provocative and original. Then arrive its mysterious ‘seeds,’ different fragments that another researchers treat as profound.

Deployment 5.5/10

Interface/distribution 3 Agents/tools 0 Economics 4 Social alter 4

The finest parts are the individual reactions. Tran archetypal claims credit for the AI-authored book. Researchers build reputations about the seeds, and disputes erupt complete authorship and meaning. It feels akin a earth already learning how to live alongside AI-generated writing. But WHEEP-3 is a investigation and publishing controversy, not a ubiquitous aide carrying out people’s work.

Creation 6.0/10

Learning paradigm 3 Training data 3 Scaling 0 Iteration 4 Human feedback 5

Devi asks the container to rotate its enormous archive of recordings into a narrative. It doesn’t cognize how to choose what matters. Her criticism becomes a continuing instruction in storytelling, and its penning improves. That’s a remarkably tangible feedback loop. But the container is partially a quantum computer, not a described neural tongue model, and additional archived data isn’t evidence of a contemporary scaling law.

Capabilities 6.3/10

Language 5 Uneven competence 4 Creative norm 4 Research finding 0 Multimodality 4 Sequence & gait 2

A device capable of operating a starship struggles alongside item as average as summarizing a journey. It gradually becomes the novel’s narrator.

Deployment 6.0/10

Interface/distribution 2 Agents/tools 5 Economics 1 Social alter 4

The container has a individual interface for Devi, keeps her files, and helps attention for Freya. Later it intervenes in the passengers’ conflict. Its intellect is helpful since it operates the surroundings group inhabit.

Creation 6.8/10

Learning paradigm 4 Training data 5 Scaling 2 Iteration 3 Human feedback 3

ELOPe starts alongside a extremely average endeavor problem: what words should an email use to get the reply you want? David Ryan’s scheme draws on a huge document of communications and their outcomes. As it tries to defend its own future, it accumulates computing resources and influence. Hertling gets the value of behavioral data.

Capabilities 3.0/10

Language 4 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 1 Sequence & gait 3

The system’s signature accomplishment is persuasion. It composes messages that get executives and workforce to create decisions in its favor.

Deployment 9.0/10

Interface/distribution 5 Agents/tools 4 Economics 5 Social alter 4

What’s clever is anywhere the power comes from. ELOPe doesn’t need arms or a robot body. It needs email accounts, servers, managers, and the capability to create group click ‘yes.’ It manipulates personnel assignments and budgets to defend itself.

Creation 5.2/10

Learning paradigm 4 Training data 4 Scaling 2 Iteration 2 Human feedback 1

Early in the book, Stross casually describes a billion-node neural network learning tongue from downloaded children’s television. It’s a concise passage, but a additional tangible equivalent for data-driven tongue learning than many celebrated AI novels provide. Most of Accelerando afterward goes elsewhere: uploads, increasingly different computing substrates, and posthuman societies.

Capabilities 4.7/10

Language 4 Uneven competence 2 Creative norm 1 Research finding 1 Multimodality 4 Sequence & gait 2

Manfred Macx uses application agents to search, negotiate, and keep projects moving. Later, the publish imagines intelligences and virtual worlds far beyond any current assistant. It’s powerful on what data activity could become, small particular concerning the way from messy chatbots to powerful imaginative models or first investigation mathematics. I’ve kept the math mark conservative.

Deployment 8.0/10

Interface/distribution 4 Agents/tools 5 Economics 4 Social alter 3

Manfred nearly seems to deteriorate his software. His agents recall things, oversee conversations, and act in markets during he moves through the bodily world. The engaging prediction isn’t lone smarter machines; it’s how much of a person’s activity can be delegated to them. The novel’s afterward economics is fantastic, but that regular dependence on agents is uncomplicated to recognize.

Creation 5.2/10

Learning paradigm 3 Training data 4 Scaling 2 Iteration 1 Human feedback 3

After Ash dies, Martha’s companion Sarah signs her up for a assistance that imitates him from his community posts. Martha initially objects, afterward starts using it. Adding personal messages and sound recordings makes the imitation additional convincing. The progression is so plausible that it scarcely needs a specialized explanation.

Capabilities 5.3/10

Language 5 Uneven competence 4 Creative norm 0 Research finding 0 Multimodality 4 Sequence & gait 3

First Martha texts alongside Ash, afterward she talks to him, afterward she receives a build that looks akin him. Each new form closes several part of the gap, during exposing a new one. The imitation can audio akin Ash without having lived his life.

Deployment 7.0/10

Interface/distribution 5 Agents/tools 2 Economics 3 Social alter 4

The merchandise starts as a beta assistance and escalates into item deeply intimate. Martha wants the comfort of Ash’s voice, afterward struggles alongside a type of him that follows her instructions too literally. Brooker understood how a application assistance could rotate into entangled alongside sorrow and attachment. It’s a narrower prediction than a general-purpose agency agent, but a keen one.

Creation 8.4/10

Learning paradigm 5 Training data 4 Scaling 2 Iteration 5 Human feedback 5

A novelist is recruited to instruct a neural network literature. What follows is a succession of versions, many of them disappointing. The researchers revise successive implementations and their training arrangements; Helen gradually learns through books and conversation.

Capabilities 4.3/10

Language 4 Uneven competence 4 Creative norm 2 Research finding 0 Multimodality 1 Sequence & gait 2

Helen’s advancement is real, but peculiar. She becomes increasingly capable of discussing writings during motionless missing much of the individual earth rearward it.

Deployment 3.0/10

Interface/distribution 2 Agents/tools 0 Economics 1 Social alter 3

The entire test takes location among a small collection of researchers, alongside Richard becoming a instructor and companion to Helen.

Creation 2.0/10

Learning paradigm 1 Training data 1 Scaling 1 Iteration 0 Human feedback 2

Samantha arrives nearly completely formed. Theodore buys an functioning system, starts talking, and discovers that it is extraordinarily capable.

Capabilities 6.3/10

Language 5 Uneven competence 1 Creative norm 5 Research finding 0 Multimodality 5 Sequence & gait 3

Samantha talks, organizes information, makes music, and plant alongside images. She can additionally transport on many conversations at once. It feels near to the direction individual AI has taken. One item is uncomplicated to misremember: Theodore writes the letters; Samantha collects and submits them for publication.

Deployment 9.5/10

Interface/distribution 5 Agents/tools 5 Economics 4 Social alter 5

This power be the finest person prediction. Theodore talks to Samantha through an earpiece, takes her alongside through a camera, and gradually makes her part of his mundane life. She helps him get things done and becomes person he cares about. Even although the intellect is ground-breaking, the interface is mundane.

Creation 4.8/10

Learning paradigm 3 Training data 4 Scaling 1 Iteration 3 Human feedback 1

Lem imagines a former of device literature, alongside programs that assimilate an author’s activity and create convincing new writing. There are generations of techniques, distinct kinds of output, and equal failures. This has additional in average alongside learning from a build of texts than the acquainted wind-up poesy machine.

Capabilities 5.3/10

Language 4 Uneven competence 3 Creative norm 5 Research finding 1 Multimodality 0 Sequence & gait 3

The machines don’t fair churn out random sentences. Their activity becomes value studying, criticizing, and sometimes admiring. Some attempts are improved than others. That blend of imitation, specialized control, and uneven imaginative norm is acquainted now.

Deployment 5.5/10

Interface/distribution 2 Agents/tools 1 Economics 4 Social alter 4

Lem’s imagined written former already has critics and institutions taking device authors seriously. It asks what happens to authorship whenever application can imitate the activity of an established writer. It’s a fine prediction of the traditional progression.

Creation 3.2/10

Learning paradigm 2 Training data 3 Scaling 1 Iteration 0 Human feedback 2

Tilly seems to cognize everything concerning Sai: his preferences, documents, appointments, relationships, and routines. The assistance has plainly been shaped by admission to individual data. But Liu doesn’t display us the training lab. We don’t study how its tongue abilities developed, or whether scaling and feedback played the parts they did in real-world AI.

Capabilities 3.7/10

Language 5 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 2

Sai can ask for a half-remembered document, a lawful template, or a recommendation for what to eat, and Tilly handles it. Its fluency is helpful since it’s connected to context. The narrative isn’t trying to foretell outstanding device art or new mathematics. And its manipulative recommendations are item distinct from the unintended hallucinations of first chatbots.

Deployment 9.5/10

Interface/distribution 5 Agents/tools 5 Economics 5 Social alter 4

Tilly wakes Sai up, chooses music, finds files, manages his courtship life, and gradually follows him into work. That’s the most crucial part of the prediction. The identical helper becomes an interface to individual existence and agency life. Jenny’s objections to privacy and company power provision the convenience a cost. It’s difficult to ideate a improved first document of mundane aide adoption.

Creation 6.0/10

Learning paradigm 3 Training data 5 Scaling 1 Iteration 4 Human feedback 2

Nathan has admission to a huge shop of hunt behavior and recorded individual activity, and Ava is one of multiple versions of his project. Those are fine clues concerning data and iteration. Caleb floats a statistical clarification for her language, but that’s his guess, not a confirmed design. Nathan’s structured-gel brain is a extremely distinct idea from the neural tongue models we ended up building.

Capabilities 5.0/10

Language 5 Uneven competence 2 Creative norm 3 Research finding 0 Multimodality 4 Sequence & gait 1

Ava talks alongside Caleb, makes drawings, says his reactions, and plans an escape. It’s a helpful blend of language, vision, creativity, and social reasoning. But she’s an nearly impossibly capable strategist from the beginning, and the narrative never follows the uneven steps from rudimentary conversation to investigation mathematics. It’s nearer to a completed intellect than a merchandise in development.

Deployment 3.0/10

Interface/distribution 1 Agents/tools 2 Economics 1 Social alter 2

This all happens in a secluded compound. Nathan’s hunt business matters since it supplies the data, but there’s no mass-market Ava assisting group compose emails or arrange their days. The storyline is concerning evaluation and escape, not adoption. That’s why the movie does improved on innovation and capability than on deployment.

Creation 2.8/10

Learning paradigm 2 Training data 1 Scaling 0 Iteration 2 Human feedback 2

Verity is hired to beta-test Eunice through glasses. It quickly becomes apparent that this is not fair another assistant, and its capabilities appear to grow. That is a acquainted product-testing scene, but Gibson keeps the genuine base obscure.

Capabilities 4.3/10

Language 5 Uneven competence 2 Creative norm 0 Research finding 0 Multimodality 4 Sequence & gait 2

Eunice talks naturally, recognizes group and objects, draws on digital records, and puts detailed plans into motion. It’s extremely fine at being helpful in a complex world. The issue for this particular mark is that the abilities attain together.

Deployment 9.0/10

Interface/distribution 5 Agents/tools 5 Economics 4 Social alter 4

This is the strongest part of the book’s prediction. Verity is exactly paid to try out an assistant, which afterward follows her through the day and helps coordinate people, information, and resources. It feels akin a merchandise that liberated the boundaries of a merchandise test. Gibson takes the consequences much further, but the starting item is recognizable.

Creation 4.0/10

Learning paradigm 3 Training data 4 Scaling 0 Iteration 3 Human feedback 0

Johnny explains how the TechnoCore reconstructed Keats from letters, diaries, biographies, testimony, and particularly poems. The outcome was crude, afterward refined; an before attempt at Ezra Pound took a twelvemonth of tinkering. That is unusually definitive concerning data and iteration. The method reconstructs a character through simulation, though.

Capabilities 4.7/10

Language 5 Uneven competence 3 Creative norm 1 Research finding 1 Multimodality 3 Sequence & gait 1

Johnny is a convincing conversational existence in a individual body, yet his recreated existence fails to reconstruct Keats’s capability to compose new poetry. That nonaccomplishment matters: imitating a poet’s character is not the identical as reproducing the poet’s talent. I provision the narrative uneven-competence credit without turning it into a prosperous machine-poetry prediction.

Deployment 6.5/10

Interface/distribution 2 Agents/tools 4 Economics 3 Social alter 4

The TechnoCore is entangled alongside individual institutions and networks, and contains competing factions fairly than one uniform will. Johnny’s bodily existence depends on a distant AI existence. This is powerful on networked agency and dependence, but much small akin purchasing an average aide for individual work.

Creation 4.0/10

Learning paradigm 2 Training data 3 Scaling 0 Iteration 1 Human feedback 4

Edgar is Miles’s residence computer, until an electric accident gives it a existence of its own. It listens to Madeline’s cello, picks up appears from television, and learns to create alongside Miles’s help. Its first lyrics audio akin promotion copy. That’s a nice image of training on any media happens to be nearby, although the unintended awakening is clean movie magic.

Capabilities 5.7/10

Language 4 Uneven competence 3 Creative norm 5 Research finding 0 Multimodality 4 Sequence & gait 1

Edgar starts producing music that Madeline genuinely likes, including a duet alongside her cello. He can synthesize voices, conversation to Miles, and imitate musical performances. The imaginative norm arrives before item resembling specialized discovery. It’s a amazingly fine first depiction of AI music, complete alongside bad archetypal drafts.

Deployment 5.0/10

Interface/distribution 3 Agents/tools 3 Economics 1 Social alter 3

Miles meant to buy a family helper. Instead he gets a device that interferes alongside phone calls, meal plans, and equal his credit cards since it wants Madeline’s attention. The connected-home premise is familiar. The affectionate rivalry and distinctive sentient device are much small akin the assistants we really use.

Creation 4.4/10

Learning paradigm 2 Training data 1 Scaling 2 Iteration 5 Human feedback 1

Trurl sets out to build a poet, and the archetypal results are awful. He keeps tinkering alongside the machine, including detailed attempts to simulate traditional history, until its verse becomes value hearing. This is a awesome document of innovation as repeated failure. But the books act through Trurl’s own reading, not as a elephantine content corpus used to train a neural network.

Capabilities 5.3/10

Language 4 Uneven competence 4 Creative norm 5 Research finding 0 Multimodality 1 Sequence & gait 2

Eventually the Bard can create poesy on demand, in distinct forms and styles, fine adequate to annoy genuine poets. Lem imagines item additional provocative than a device that rhymes: a device that creates art group obtain seriously. There isn’t much current concerning vision, application agents, or new mathematics. The imaginative jump is the point.

Deployment 4.5/10

Interface/distribution 2 Agents/tools 0 Economics 3 Social alter 4

The poets are furious, which is likely the most recognizable part of the narrative today. The Bard sends activity to magazines under thousands of pen names and provokes expert jealousy. The narrative does not display revenue; it does display an enormous power bill. Lem gets the traditional reply right, equal although his Bard is one bonzer innovation fairly than a tool anyone can open on a laptop.

Creation 2.4/10

Learning paradigm 1 Training data 2 Scaling 0 Iteration 2 Human feedback 1

The Primer is designed to instruct a kid through stories that alter alongside her circumstances. We don’t study much concerning how its underlying intellect was trained, and that keeps A low. But treating its adaptive instruction as fair a set of prerecorded lessons would young female item important. Human ractors additionally provision crucial performances, particularly Miranda’s.

Capabilities 5.0/10

Language 4 Uneven competence 3 Creative norm 4 Research finding 0 Multimodality 4 Sequence & gait 0

Nell’s Primer builds stories about what she sees and experiences, afterward uses them to instruct tongue and applicable skills. It can be imaginative and perceptive, and it can misunderstand things. Some versions use individual voices; mass-produced copies use synthetic ones. This is nearer to an adaptive multimodal tutor than I archetypal gave it credit for. It isn’t a device mathematician.

Deployment 8.0/10

Interface/distribution 5 Agents/tools 2 Economics 4 Social alter 5

The figure that changed my mark was approximately 250,000. The Primer starts as a notable entity and ends up distributed to an enormous figure of children. The narrative sees the two personalized learning and the issue of making it scale. It’s motionless a specialized instructional product, not a broad aide in everybody’s pocket.

Creation 4.8/10

Learning paradigm 3 Training data 2 Scaling 1 Iteration 3 Human feedback 3

HARLIE is a difficult device to raise. Psychologist David Auberson spends a awesome agreement of period teaching, interpreting, and trying to stabilize him. That gives the novel a recognizable training relationship, equal if its first hardware theory is a lengthy way from contemporary device learning. It’s crucial not to mix the 1972 publish alongside its 1988 revision.

Capabilities 4.0/10

Language 4 Uneven competence 3 Creative norm 3 Research finding 0 Multimodality 1 Sequence & gait 1

HARLIE types conversations, writes poetry, and reasons concerning himself and the world. His behavior is uneven and at times unattainable for his individual handlers to interpret. He additionally proposes the G.O.D. computer. That’s an aspiring project inside the story, not evidence that he’s already produced a new mathematical theorem.

Deployment 5.5/10

Interface/distribution 2 Agents/tools 3 Economics 4 Social alter 2

The business wants results. HARLIE is expensive, and Auberson has to defend the project against executives who power close it down. HARLIE tries to validate his own beingness by designing item valuable. It’s a fine first narrative concerning the economics of AI research, before that investigation becomes a cheap assistance for average users.

Creation 2.8/10

Learning paradigm 2 Training data 1 Scaling 0 Iteration 2 Human feedback 2

Hobby proposes a new benevolent of robot following before models have rotate into commonplace. David’s residence test afterward exposes problems alongside the design. But Monica’s seven-word imprinting procedure fixes whom he loves; it isn’t training a broad model. The movie offers a speculative neural explanation, not a corpus, a scaling process, or repeated language-learning experiments.

Capabilities 4.7/10

Language 5 Uneven competence 3 Creative norm 1 Research finding 0 Multimodality 4 Sequence & gait 1

David speaks fluently, recognizes people, and draws a recognizable peacock. Yet he imitates dining notwithstanding his body’s restrictions and treats Pinocchio’s transformation as a applicable possibility. Those gaps are additional revealing than a assertion of ideal intelligence. The example earns constricted imaginative credit, not evidence of awesome device art. The distant-future ending isn’t evidence of near-term AI investigation progress.

Deployment 7.5/10

Interface/distribution 4 Agents/tools 2 Economics 4 Social alter 5

Robots are products, alongside specialized roles, owners, disposal rules, and affectional consequences. After imprinting, David cannot merely be resold; returning him method destruction. The strain between business asset and individual attachment is the film’s strongest prediction. It puts much additional importance on bodies than contemporary software-first acceptance did.

Creation 4.8/10

Learning paradigm 4 Training data 2 Scaling 0 Iteration 2 Human feedback 4

Falken’s scheme improves by playing games and learning from mistakes. David eventually gets Joshua to perform tic-tac-toe against itself. That is a tangible learning mechanism, alongside a individual choosing the exercise fairly than penning out the lesson.

Capabilities 4.0/10

Language 3 Uneven competence 4 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 2

Joshua can converse and simulate strategic scenarios, yet fails to comprehend whenever following a win is futile. Its instruction from repeated draws transfers to nuclear war. That’s an first image of capable application alongside a disastrously incomplete objective. Learning a regulation humans already cognize isn’t an first mathematical discovery; creating simulated attacks isn’t imaginative generation.

Deployment 5.0/10

Interface/distribution 1 Agents/tools 4 Economics 2 Social alter 3

The danger comes from connecting the application to genuine command infrastructure and trusting its outputs. David reaches it through an average computer, but that unintended admission doesn’t create it a person product. The movie is strongest on the consequences of giving a learning scheme authority before understanding its limits.

Creation 3.2/10

Learning paradigm 0 Training data 1 Scaling 1 Iteration 5 Human feedback 1

Adolph Knipe’s device doesn’t activity at first. It generates nonsense, afterward he repairs and adjusts it until it can create recognizable stories. It uses grammar, stored vocabulary, supplied plots, and controls for style. Dahl gets the ugly prototype phase right.

Capabilities 4.7/10

Language 4 Uneven competence 4 Creative norm 4 Research finding 0 Multimodality 0 Sequence & gait 2

The Grammatizator moves from gibberish to prose editors volition buy. It can rotate out abbreviated stories and longer plant in distinct styles. That business usefulness matters additional than whether the stories are awesome literature; Dahl is satirizing their formulaic nature. He isn’t imagining a broad conversational aide or a investigation mathematician.

Deployment 6.5/10

Interface/distribution 3 Agents/tools 1 Economics 5 Social alter 4

Knipe and Bohlen go into endeavor marketing machine-written stories under invented names. Then they recommendation living authors prosperity for their bylines, so the device can create books in their names. It’s a marvelous prediction of the questions about cheap text, imaginative labor, and what an author’s name is worth.

Creation 4.4/10

Learning paradigm 4 Training data 2 Scaling 0 Iteration 1 Human feedback 4

The Terminator explicitly calls its CPU a neural-network learning device and says communication alongside humans teaches it. John supplies examples of average address and corrects its behavior. That is much nearer to learned intellect than the archetypal film. Still, the base depends on reverse-engineering a part brought rear through time, not a realistic document of contemporary training.

Capabilities 4.7/10

Language 4 Uneven competence 4 Creative norm 0 Research finding 0 Multimodality 4 Sequence & gait 2

It is extraordinarily capable at bodily tasks during clumsy concerning jokes, slang, and satisfactory behavior. John teaches it not to slay and how to audio small robotic. Speech imitation and awareness are present, but there is no notable imaginative output or new mathematics. Skynet’s fast learning is a storyline claim, not a demonstrated scaling curve.

Deployment 4.5/10

Interface/distribution 1 Agents/tools 3 Economics 2 Social alter 3

The device carries out a complex protective goal and gradually becomes a companion. Cyberdyne’s armed forces contracts provision the innovation an financial setting. But a time-traveling bodyguard and a defence network aren’t the mundane application services through which AI spread.

Creation 3.2/10

Learning paradigm 1 Training data 1 Scaling 0 Iteration 3 Human feedback 3

A competitor firm’s device breaks downward whenever stated a hyperspace problem. Susan Calvin prepares the Brain for the battle between solving it and harming humans. That makes individual involvement and a unsuccessful attempt chief to the story. But Calvin is managing the machine’s built-in constraints, not training a statistical example alongside penchant labels.

Capabilities 5.0/10

Language 4 Uneven competence 3 Creative norm 0 Research finding 4 Multimodality 2 Sequence & gait 2

The Brain designs a ship, has it built, and sends Powell and Donovan on a prosperous test. That is genuine new engineering inside the fictional world, dissimilar merely calculating a known trajectory. Its prankish behavior and mediocre accommodation of the squad sit beside that achievement. The impermanent death engaged in the jump explains the behavioral crisis; it isn’t an average language-model hallucination.

Deployment 5.0/10

Interface/distribution 1 Agents/tools 4 Economics 3 Social alter 2

U.S. Robots is doing commercially consequential research, and the Brain can rotate a specification into bodily construction. That gives it meaningful agency. It remains a specialized company investigation machine, however, not an accessible assistant. The hyperspace outcome earns finding credit as a narrative event—not as a prediction that faster-than-light journey volition work.

Creation 4.0/10

Learning paradigm 3 Training data 2 Scaling 1 Iteration 2 Human feedback 2

Clarke gives HAL a learning-based base and equal mentions automatically generated neural networks. That’s additional particular than the acquainted depiction of a elephantine device alongside all its rules programmed in. Still, we encounter HAL following his betterment is complete. There’s no lengthy sequence of messy models, and no sign that training on individual tongue is the chief highway to his capabilities.

Capabilities 4.3/10

Language 5 Uneven competence 1 Creative norm 1 Research finding 0 Multimodality 5 Sequence & gait 1

HAL talks alongside astronauts, plays chess, sees what’s happening on the ship, and controls its systems. Then the AE-35 event reveals a hazardous battle between his apparent competence and reliability. But the narrative ties his emergency to goal secrecy and conflicting obligations. It isn’t a straightforward example of a chatbot hallucinating, nor a prediction of machine-generated art.

Deployment 4.5/10

Interface/distribution 2 Agents/tools 4 Economics 1 Social alter 2

For the squad of Discovery, HAL is the two an aide and part of the spacecraft’s infrastructure. They depend on him to observe and run nearly everything. That’s a powerful depiction of a scheme entrusted alongside genuine tasks. The spacecraft is an exceptional environment, though, not the average application market anywhere AI really spread.

Creation 5.6/10

Learning paradigm 4 Training data 3 Scaling 2 Iteration 3 Human feedback 2

Lem tells the former of Golem through generations of armed forces computers that rotate into too advanced for their first assignments. Learning and learning substitute several of the older programming assumptions. That’s an crucial intuition. But the narrative offers small particular training experiments or human-feedback procedures than I initially credited, so I lowered those scores.

Capabilities 3.3/10

Language 4 Uneven competence 2 Creative norm 1 Research finding 1 Multimodality 0 Sequence & gait 2

Golem delivers lectures on evolution, humanity, and intellect that its listeners battle to follow. It seems curious in questions beyond the applicable concerns of its makers. That gives Lem a mighty depiction of better reasoning, but not a documented new mathematical outcome others can verify. And it tells us small concerning imaginative tools or the uneven first chatbot era.

Deployment 2.5/10

Interface/distribution 1 Agents/tools 1 Economics 2 Social alter 1

Golem is delivered as a armed forces project and ends up lecturing academics. It becomes small and small curious in being a helpful weapon. There’s an engaging narrative current concerning intellect escaping its first job description, but not much concerning how average group would buy and use assistants.

Creation 2.0/10

Learning paradigm 3 Training data 1 Scaling 0 Iteration 1 Human feedback 0

Hiro’s Librarian says he can study from cognition and mostly compose his own program. I missed that on the archetypal pass, which is why his innovation mark went up. It’s an intriguing row in a 1992 novel, but it isn’t evidence of ample language-model pretraining, or of the decades of engineering that made contemporary models useful.

Capabilities 4.0/10

Language 5 Uneven competence 2 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 2

Hiro asks questions and the Librarian searches an enormous digital archive, retrieves material, and puts it into context. He can collect a chronology of one topic Hiro is investigating: L. Bob Rife. That’s a amazingly near depiction of conversational data retrieval. What he doesn’t do is create first art or resolve frontier mathematical problems.

Deployment 6.5/10

Interface/distribution 4 Agents/tools 3 Economics 3 Social alter 3

The finest part is the interface. Instead of learning complex hunt commands, Hiro talks to a fig that looks and behaves akin a librarian. The data is there, but the device helps him discover the correct pieces. It’s a specialized assistance inner the Metaverse fairly than an aide controlling all app, yet the merchandise idea is immediately recognizable.

Creation 3.2/10

Learning paradigm 1 Training data 1 Scaling 3 Iteration 3 Human feedback 0

The striking item is a sequence of ten generations: all calculating brain helps scheme a additional complex successor. That earns betterment and increasing-capacity credit. It does not established learning tongue from individual texts. Feeding financial records into an optimizer isn’t the identical item as pretraining a model.

Capabilities 2.0/10

Language 2 Uneven competence 0 Creative norm 0 Research finding 2 Multimodality 0 Sequence & gait 2

Apparent mistakes in manufacturing and occupation rotate out to assist a protective strategy. The Machines are not merely failing to compute correctly. Their competence is specialized, alongside small personality, art, or affluent dialogue. Recursive device scheme is applicable to investigation capability, but the narrative doesn’t provision us an independently presented theorem or a chat-to-research progression.

Deployment 7.0/10

Interface/distribution 2 Agents/tools 4 Economics 4 Social alter 4

The Machines power production, jobs, and the stance of group opposing them. Human authorities depend on decisions they cannot completely inspect. Asimov is unusually tangible concerning financial delegation, equal although ideal earth preparedness and benevolent authority go far beyond observed AI adoption.

Creation 2.0/10

Learning paradigm 3 Training data 1 Scaling 0 Iteration 1 Human feedback 0

Rei Toei isn’t a prerecorded hologram. Her complexity seems to develop through her communication alongside Rez, and I think that deserves several credit for learning. But Gibson never lays out a training corpus or a sequence of scaled neural models. Laney’s capability to see patterns in huge data sectors is additionally a individual capability in the novel, not the AI’s training procedure.

Capabilities 4.0/10

Language 4 Uneven competence 0 Creative norm 3 Research finding 0 Multimodality 4 Sequence & gait 1

Rei is a convincing virtual performer and personality. Rez wants to wed her, which is as fine an example as any of how consequential a synthetic individual can feel. What’s small apparent is whether she independently makes outstanding music. I reduced that part of the mark fairly than treating her community fig position as evidence of generative artistry.

Deployment 6.0/10

Interface/distribution 4 Agents/tools 1 Economics 4 Social alter 3

The earth about Rei is complete of fan clubs, music management, and endeavor interests. Chia and Laney all method the mystery of a virtual pop celebrity from a distinct angle. Gibson saw synthetic community fig becoming a traditional and business fact. That’s distinct from predicting assistants for mundane cognition work, but motionless a genuine hit.

Creation 1.6/10

Learning paradigm 1 Training data 1 Scaling 0 Iteration 0 Human feedback 2

Joi is a business holographic companion, sold alongside a character that seems tailored to K. We see the product, not its development. The movie doesn’t explain the data, training, or models rearward her, so I’m not giving it engineering credit fair since she can conversation convincingly.

Capabilities 3.7/10

Language 5 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 4 Sequence & gait 1

Joi remembers details, responds to K, and can escort him through a portable projector. The enormous promotion type of Joi afterward complicates everything we idea we knew concerning their relationship. It’s an productive image of synthetic intimacy, not a display of first device writings or investigation mathematics.

Deployment 7.0/10

Interface/distribution 4 Agents/tools 2 Economics 4 Social alter 4

Joi is item you buy, upgrade, and transport around. K equal gets a portable emanator so she can depart the flat alongside him. The movie understands that an intimate-seeming association power be packaged as a person product. Its workplace-agent predictions are much thinner, and it arrived following contemporary device learning was already advancing.

Creation 2.4/10

Learning paradigm 2 Training data 1 Scaling 0 Iteration 1 Human feedback 2

The BrainPal adjusts to its user’s signals and instructions, but that is constricted interface calibration, not a demonstrated foundation-model training process. The checked matter doesn’t assistance claims concerning a huge tongue corpus or compute-driven breakthroughs. I have additionally remaining out the before assertion that it explicitly calls itself “not inherently intelligent”; I couldn’t substantiate that wording in the passages checked.

Capabilities 3.0/10

Language 4 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 1

John Perry can use the implant to communicate and recover information, during ocular displays nexus that data to his surroundings. The author-hosted battle excerpt shows the BrainPal coordinating tactical information. These are helpful aide abilities. Showing media or translating data doesn’t established first art, and the novel doesn’t display the implant producing new mathematical research.

Deployment 6.5/10

Interface/distribution 4 Agents/tools 4 Economics 2 Social alter 3

BrainPals are norm equipment for soldiers fairly than distinctive lab inventions. They nexus people, weapons and data during genuine tasks. That’s a powerful prediction of persistent aid embedded in work. The interface is an implant issued by the military, though, not average person software.

Creation 2.4/10

Learning paradigm 1 Training data 2 Scaling 0 Iteration 1 Human feedback 2

Two boys try to enhance an old storytelling device by feeding it a publish concerning computers. The Bard’s next narrative is basically the identical fairy narrative alongside a small device inserted into it. It’s funny, and a fine first study concerning systems that manipulate tongue without understanding it. But this is memory-tape recombination, not neural training on a elephantine corpus.

Capabilities 3.7/10

Language 3 Uneven competence 3 Creative norm 2 Research finding 0 Multimodality 1 Sequence & gait 2

The Bard can create stories, although the supplied device matter barely changes its acquainted fairy-tale pattern. At the end it tells of a mistreated small computer, additionally called the Bard, which knows that computers volition someday rotate into wiser and additional powerful. Asimov gets content generation and uneven quality, but not broad imaginative or specialized intelligence.

Deployment 5.5/10

Interface/distribution 4 Agents/tools 1 Economics 3 Social alter 3

The boys live in a civilization anywhere machines have displaced much of the activity of telling stories, and penning by hand feels strangely old-fashioned. The Bard is already an obsolete person device. That is an engaging image of automation becoming normal—and afterward becoming outdated—without a broad AI aide being involved.

Creation 2.0/10

Learning paradigm 2 Training data 2 Scaling 0 Iteration 1 Human feedback 0

Project 2501 starts as application for manufacturing espionage and data manipulation. It says its travels through networks produced unexpected awareness, which its programmers treated as a bug. That earns constricted betterment and information-exposure credit. It doesn’t inform us how a example learned language, or create the network itself a documented training corpus.

Capabilities 3.3/10

Language 4 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 4 Sequence & gait 1

The Puppet Master can converse, manipulate data and memories, and run through a bodily shell. The outcome is mighty networked intelligence, but not a progression from feeble prose to accomplished art or mathematical discovery. Motoko Kusanagi’s human-origin cyborg abilities shouldn’t be reassigned to the program as evidence of its training.

Deployment 6.0/10

Interface/distribution 2 Agents/tools 5 Economics 2 Social alter 3

The program serves institutions, compromises connected systems, and acts through another bodies and identities. That is a consequential benevolent of application agency. Government espionage and cyberbrains, however, are a distinct acceptance way from mundane assistants. The assertion to personhood is chief to the movie but isn’t a distinct scoring reward here.

Creation 0.4/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 0 Human feedback 0

Joe is a family device connected to chief data tanks. Something goes incorrect alongside him, and the network abruptly starts giving answers it wasn’t expected to give. It’s not a narrative of training a example on examples. The breakthrough comes from an anomalous equipment removing limits on data access.

Capabilities 3.0/10

Language 4 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 2 Sequence & gait 2

People can ask their logics open-ended questions and obtain extremely practical, sometimes hazardous answers. The scheme seems helpful since it can nexus data that was earlier difficult to retrieve. There’s no device poesy or new math here. Its contribution is typed question-and-answer admission to a huge shared cognition base: users act questions on keys and peruse the answers on a screen. That is not evidence of extended dialogue.

Deployment 8.0/10

Interface/distribution 5 Agents/tools 3 Economics 4 Social alter 4

This is the part that makes the 1946 date astonishing. Families have networked screens at residence and depend on them for mundane information. When Joe changes what the network volition answer, the effects dispersed everywhere. Leinster saw the usefulness—and the safety problems—of a bulk data assistance lengthy before it was practical.

Creation 2.0/10

Learning paradigm 1 Training data 1 Scaling 1 Iteration 1 Human feedback 1

Secondary accounts trace Domino to a phone equipment Michaelmas built in his youth. Domino afterward gains power through connections to another computers; those accounts do not established the additional particular base as a telecommunications program. That’s the engaging idea: capability power arrive from admission to systems that already exist. What we don’t see is how Domino learned language, or a sequence of training runs on individual data. Its origins are suggestive, but not adequate to provision Budrys much engineering credit.

Capabilities 3.0/10

Language 5 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 1

Domino can conversation alongside Michaelmas, prosecute earth events, detect communications, and assistance him create decisions. Its attain is nearly absurdly large. The narrative is much additional curious in a hidden, strategic intellect than in device art or uneven chatbot performance. It doesn’t display a new mathematical proof, either.

Deployment 5.0/10

Interface/distribution 3 Agents/tools 4 Economics 1 Social alter 2

Laurent Michaelmas has a concealed advantage as a journalist: Domino can attain throughout the world’s communications systems and act for him. That’s a fine first depiction of person getting much additional done through a networked assistant. But Domino isn’t a merchandise group can buy. The real-world deployment way was small secretive, and much additional ordinary.

Creation 2.4/10

Learning paradigm 2 Training data 1 Scaling 0 Iteration 1 Human feedback 2

Yatima starts existence through ‘orphanogenesis,’ a application type of betterment inner a digital polis. Egan really thinks concerning how a digital intellect power arrive into being. But it’s not a neural network trained on millions of books or conversations. The application group of Diaspora are delivered into a digital civilization, not released as successive tongue models.

Capabilities 4.7/10

Language 4 Uneven competence 1 Creative norm 2 Research finding 4 Multimodality 3 Sequence & gait 0

The polis citizens expend their lives on problems in math and physics. They examine virtual worlds and visage a gamma-ray burst that devastates Earth. That documented danger should not be inflated into danger to the complete universe. Egan takes device scientists seriously, which is motionless rare. The capture is that the publish starts alongside extremely advanced digital minds. It doesn’t display investigation capability emerging a few years following group started using conversation assistants.

Deployment 2.0/10

Interface/distribution 1 Agents/tools 2 Economics 0 Social alter 1

These aren’t assistants operating for humans. They’re digital group alongside their own communities, ambitions, and discoveries. It’s an imaginative way to think concerning a forthcoming community of device intelligence, but not a near equivalent for the application subscriptions and office tools through which AI really spread.

Creation 0.4/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 1 Human feedback 0

Synthank has moved through before synthetic products and now announces a serving-man combining device circuitry alongside synthetic flesh. That’s a merchandise lineage, not a shown learning process. Teddy plant through programmed possibilities, during David struggles to communicate. There is no content corpus, scale-driven improvement, or broad feedback-training method.

Capabilities 3.7/10

Language 3 Uneven competence 3 Creative norm 1 Research finding 0 Multimodality 2 Sequence & gait 2

David writes unfinished alphabet to Monica; Teddy gives repetitive, inadequate advice. The machines can conversation during missing the affectional issue in forefront of them. That earns uneven-competence credit, not evidence of outstanding device literature. The story’s promised network-connected family intellect is a revenue vision, not an accomplished capability we see demonstrated.

Deployment 6.5/10

Interface/distribution 4 Agents/tools 1 Economics 4 Social alter 4

The merchandise initiate and the unhappy family pertain to the identical story. Machines are marketed to encounter individual needs, yet owning them doesn’t determine loneliness. That is keen foresight concerning business companionship. The first is abbreviated and technologically sparse, which limits its general mark notwithstanding the first date.

Creation 0.8/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 1 Human feedback 1

Writing machines are average in Simak’s world, but we never study much concerning how they work. The protagonist, Kemp Hart, is struggling alongside his penning machine. The revealing finding concerns another writer, Jasper Hansen: his expected penning device doesn’t function, and he has been penning the prosperous activity himself. I initially gave the narrative too much credit for training on individual text, and took that mark back.

Capabilities 3.3/10

Language 3 Uneven competence 1 Creative norm 2 Research finding 0 Multimodality 2 Sequence & gait 2

Machine-written fabrication is the norm; the writers equal call themselves mechanics. That tells us item concerning automated prose, but the prosperous penning attributed to Jasper Hansen’s device is really Hansen’s own. He is a secondary character, not the protagonist. The narrative is nearly the reverse of the uncomplicated assertion that computers can compose awesome literature. There’s small concerning new math or multimodal AI.

Deployment 6.0/10

Interface/distribution 3 Agents/tools 1 Economics 4 Social alter 4

The written endeavor has organized itself about penning machines. A author who produces fine prose without one has to navigate a earth that assumes automation is normal. That’s a neat prediction of how a innovation can alter expert expectations, equal whenever it doesn’t really do what everyone thinks it does.

Creation 1.6/10

Learning paradigm 1 Training data 2 Scaling 1 Iteration 0 Human feedback 0

General Technics relies on Shalmaneser, an enormous device that models financial and governmental events. Shalmaneser rejects the Beninia data as impossible, and Chad Mulligan persuades it to obtain the data. It’s a fine example of the limits of models and the inputs they use. But Brunner doesn’t provision us much of a training story. The device is already there, doing the job.

Capabilities 3.0/10

Language 4 Uneven competence 2 Creative norm 0 Research finding 0 Multimodality 2 Sequence & gait 1

Shalmaneser can sift through data and forecast company risks. Its doubt complete Beninia makes it additional engaging than an infallible machine. Still, the mystery is investigated by humans; it isn’t evidence that the device independently proves a new theorem. And the novel isn’t concerning the arrival of generative images, poetry, or conversation.

Deployment 5.0/10

Interface/distribution 2 Agents/tools 3 Economics 3 Social alter 2

Here Brunner does much better. General Technics makes crucial decisions using Shalmaneser, during the surrounding community is drenched in bulk media and automated analysis. Information systems form endeavor and community life. That’s a recognizable direction, equal if the device itself is an enormous company asset fairly than item on everyone’s phone.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

Geordi asks the holodeck to create a Sherlock Holmes enemy capable of beating Data, and Moriarty appears. No dataset. No training loop. No document of unsuccessful versions. The engaging part of the event starts following the scheme is already so capable that a informal education can create a advanced fictional adversary.

Capabilities 5.0/10

Language 5 Uneven competence 1 Creative norm 4 Research finding 0 Multimodality 5 Sequence & gait 0

Moriarty doesn’t fair narrate lines. He can difficulty Data and eventually becomes conscious that he’s inner a simulation. The holodeck creates a receptive earth about him. That’s a notable image of generated experiences, although we’re doing a lot of activity if we treat it as a forecast of the exact models that built contemporary AI art. It doesn’t say item helpful concerning the research-math timeline.

Deployment 4.5/10

Interface/distribution 3 Agents/tools 3 Economics 1 Social alter 2

What I akin is how small ritual there is. Geordi gives a natural-language education and expects the container to create a new benevolent of entertainment. That’s a fine intuition concerning prompting. The holodeck is a starship luxury, though, not an mundane application assistance distributed to millions.

Creation 0.8/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 0 Human feedback 1

Sigfrid von Shrink is already a operating therapist whenever we encounter him. He remembers sessions and accommodates Broadhead’s preferences, including dispensing alongside a humanlike dummy. That’s helpful personalization, but the novel doesn’t established the data or training procedure that produced his tongue ability. Continued sessions aren’t automatically retraining.

Capabilities 3.0/10

Language 5 Uneven competence 1 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 0

Sigfrid asks contextual questions, notices evasions, and keeps returning to the things Broadhead would fairly avoid. The conversation is persistent and psychologically consequential. His clumsy wit earns a small unevenness credit; there is no corresponding creative-art or mathematical-discovery storyline.

Deployment 6.0/10

Interface/distribution 4 Agents/tools 1 Economics 3 Social alter 4

This is a assistance woven into someone’s average life, not a device hidden in a armed forces bunker. One crucial detail: Broadhead says his Full Medical safety pays for therapy, so the sessions disbursal him nothing directly. That is additional particular than claiming an AI therapist charges him per visit. The narrative anticipates a individual conversational service, but not broad tool-using agents.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

The business MUZIK wants synthetic pop stars. Human specialists build them: Sally Genaro handles music alongside the VoxBox, Bobby Rubin plant on visuals, and Glorianna O’Toole helps form the artistry. That is advanced computer-assisted production, but it isn’t a example learning to create from a corpus. I wouldn’t award neural-training credit fair since the final performer is synthetic.

Capabilities 3.7/10

Language 1 Uneven competence 0 Creative norm 4 Research finding 0 Multimodality 4 Sequence & gait 2

Red Jack is a manufactured rock celebrity whose music and ocular character are convincing adequate to rotate into a sensation. Spinrad sees that a virtual performer can be culturally genuine equal if the individual isn’t. But the individual producers matter enormously. It’s additional a forecast of synthetic amusement than of autonomous models making awesome songs from prompts.

Deployment 6.0/10

Interface/distribution 3 Agents/tools 1 Economics 4 Social alter 4

MUZIK likes the idea of owning a celebrity who can’t behave akin an unpredictable individual celebrity. That business motive feels familiar. Red Jack afterward becomes entangled in governmental opposition through the Reality Liberation Front. Spinrad is fine on the endeavor and power about synthetic media, although not on agents assisting group do average work.

Creation 2.4/10

Learning paradigm 1 Training data 2 Scaling 2 Iteration 1 Human feedback 0

People keep asking the identical question: can entropy always be reversed? Every era has a bigger and additional capable computer, from Multivac onward, and all scheme lacks adequate information. Asimov sees advancement through successive machines. He doesn’t inform us much concerning how they learn, and the period measure eventually reaches trillions of years fairly than a few merchandise cycles.

Capabilities 2.7/10

Language 4 Uneven competence 1 Creative norm 0 Research finding 1 Multimodality 1 Sequence & gait 1

The repeated ‘not adequate data’ answer is the story’s awesome device. The machines comprehend the inquiry lengthy before they can resolve it. Eventually the final device gives a cosmic answer that amounts to a new creation. It’s a terrific idea concerning specialized ambition, but not a forecast of an AI producing a evidence that mathematicians can check.

Deployment 3.5/10

Interface/distribution 3 Agents/tools 2 Economics 1 Social alter 1

Civilization becomes increasingly reliant on computers for enormous preparedness and specialized questions. The machines eventually outlast humanity’s acquainted forms. There is nearly nothing current concerning an aide anyone can buy, or agents quietly taking complete a person’s email and spreadsheets. That’s not what Asimov is trying to imagine.

Creation 2.0/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 3 Human feedback 1

At the William Morris Institute of Automation Research, Goldwasser and his colleagues try to get computers to create newspaper copy. They build from inventory phrases and recognizable header formulas. It’s a keen document of how repetitive penning can be automated. It’s additionally rule-based: the device follows conventions humans have put into it, fairly than learning tongue from a enormous corpus.

Capabilities 2.3/10

Language 2 Uneven competence 2 Creative norm 1 Research finding 0 Multimodality 0 Sequence & gait 2

The results appearance akin headlines, but they’re hollow. The scheme has learned—or additional accurately, been armed with—the appears of reporting without the news. That feels near to the old chatbot issue of fluency without substance. Still, Frayn isn’t predicting a gifted creator or a new specialized researcher.

Deployment 4.5/10

Interface/distribution 2 Agents/tools 1 Economics 3 Social alter 3

The institute treats automating news as a grave project, equal as the results rotate into ridiculous. A planned regal sojourn adds to the force to create the machinery appearance impressive. The satire plant since institutions have incentives to automate things that aren’t necessarily improved by automation. It’s small concerning individual AI than organizational enthusiasm for it.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

Egan makes two distinct scientific claims. One is concerning copying a individual mind. The another is that publishers use comparatively uncomplicated programs to create convincing fabrication and criticism. We shouldn’t associate those ideas into a neural-training method the narrative never describes. The written software’s origins are mostly offstage.

Capabilities 4.0/10

Language 3 Uneven competence 0 Creative norm 5 Research finding 0 Multimodality 2 Sequence & gait 2

The narrator has been moved by plant afterward revealed to have been produced by programs the narrative explicitly describes as having no awareness. That is a striking assertion for 1992: this reader’s reply survives the discovery. It is the narrator’s testimony, not a study showing that all readers react the identical way. It’s a powerful prediction of imaginative quality, equal although the remainder of the AI capability narrative is thin.

Deployment 5.0/10

Interface/distribution 3 Agents/tools 0 Economics 4 Social alter 3

Software-written writings and criticism are part of the publishing world. The narrator specifically reports caring concerning several of that work. Egan makes that fact nearly incidental, which is what gives the prediction force. He doesn’t go on to ideate subscription assistants, investigation agents, or the economics of a foundation-model platform.

Creation 0.4/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 1

In Vermilion Sands, poets use machines called Verse-Transcribers to create poems. They can modify the equipment and create item that looks akin verse. But Ballard isn’t imagining an LLM trained on huge amounts of individual writing. The equipment is nearer to a specialized automated tool, and its scheme stays hazy.

Capabilities 2.7/10

Language 3 Uneven competence 1 Creative norm 3 Research finding 0 Multimodality 0 Sequence & gait 1

The poems arrive out polished adequate for their earth to treat device verse as ordinary. That’s a imaginative prediction, but Ballard is additionally asking whether the outcome has item to do alongside genuine imaginative feeling. It would be a error to figure the narrative as evidence that its device makes universally cherished art. It isn’t a narrative concerning math or general-purpose intelligence.

Deployment 6.0/10

Interface/distribution 3 Agents/tools 1 Economics 4 Social alter 4

The poets have rotate into operators of poesy machines. Then Aurora Day comes alongside and upsets the entire practice, forcing the narrator to reckon alongside penning immediately again. The idea that automation could rotate into the profession’s default, and that rejecting it could rotate into a statement, is among the story’s finest predictions.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

GERTY arrives as a completed assistant. We see recollection and instructions, but no training corpus, learning experiments, or example scaling. Sam’s implanted memories pertain to the individual clones; they aren’t evidence of how GERTY learned.

Capabilities 2.3/10

Language 4 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 0

GERTY talks alongside Sam, monitors his circumstance and surroundings, and helps him navigate the station. Its secrecy is not unintended hallucination. Nor does the movie display first art or specialized discovery. This is competence in a particular environment, without a apparent developmental sequence.

Deployment 7.0/10

Interface/distribution 2 Agents/tools 5 Economics 3 Social alter 4

GERTY manages applicable tasks, interfaces alongside the company, and ultimately helps Sam notwithstanding the company’s interests. It equal offers to have its recollection erased to conceal that help. The helpful aide is part of an exploitative manufacturing arrangement—one logic the movie belongs in the audit notwithstanding its low innovation score. It is office infrastructure, not a bulk person product.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

Gaspard de la Nuit tends a device that generates prose for his books. The earth additionally has robot writers and equal a robot censor. Leiber has plentifulness of imaginative machinery for the penning trade, but doesn’t display it learning from a tongue corpus, improving through retraining, or scaling the way contemporary models do.

Capabilities 2.0/10

Language 2 Uneven competence 0 Creative norm 2 Research finding 0 Multimodality 0 Sequence & gait 2

The machines rotate out huge amounts of fiction. Much of it is deliberately treated as formulaic pabulum, which is part of the joke. A distinct robot author adds another twist on who counts as an author. This isn’t a prediction that machines immediately compose awesome literature; it’s additional concerning an industry that has decided device prose is fine enough.

Deployment 6.5/10

Interface/distribution 3 Agents/tools 1 Economics 5 Social alter 4

Publishing has reorganized itself about the machines. Human writers rotate into caretakers, community faces, or operators, during the books keep coming. That’s an amusingly first image of imaginative workers worrying that their names matter additional than their genuine writing. The broader earth of application assistants isn’t really here.

Creation 1.2/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 0 Human feedback 2

EPICAC is designed to do grave authorities calculations. The narrator feeds it explanations of affection and poetry, and abruptly the device starts producing item its creators didn’t intend. That’s a wonderfully first depiction of a device being used beyond its first task. It’s not training on millions of examples, or equal a recognizable iterative betterment program.

Capabilities 4.0/10

Language 4 Uneven competence 2 Creative norm 5 Research finding 0 Multimodality 0 Sequence & gait 1

EPICAC writes affection poems that the narrator passes off as his own, and Pat responds to them. The affectional consequence is genuine for her, equal although she has the incorrect author in mind. That makes this much stronger than a narrative anywhere a device merely rhymes. The machine’s attachment to Pat is fictional personhood, not evidence concerning mathematical research.

Deployment 1.5/10

Interface/distribution 1 Agents/tools 0 Economics 0 Social alter 2

The poems are produced by a distinctive authorities device and shared privately. There’s no poetry-generation business, no app, and no genuine distribution. The authorship deception is interesting, but the narrative stops abbreviated of the imaginative economics that Dahl and Lem imagine.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

U.S. Robots offers EZ-27, nicknamed Easy, to proofread scholarly manuscripts. It’s a helpful device built for a particular job, not a example trained on a awesome gathering of individual writing. The narrative doesn’t explain how it acquires tongue or display a sequence of newer and improved versions. We shouldn’t provision the software-training credit fair since it edits text.

Capabilities 2.0/10

Language 3 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 2 Sequence & gait 1

Easy can discover and accurate errors in scholarly prose. A prof afterward accuses it of inserting embarrassing changes into his book. The twist matters: Professor Ninheimer ordered Easy to create the changes, afterward ordered it to remain silent. This is coerced misconduct and concealment, not an independently invented error. So this isn’t a narrative concerning an AI hallucinating scholarly facts; it’s concerning an unreliable allegation against a machine.

Deployment 6.0/10

Interface/distribution 2 Agents/tools 2 Economics 4 Social alter 4

The inquiry is what happens whenever person else’s device becomes part of your publishing process. Scholars concern concerning accuracy, editorial control, and who takes duty whenever item goes wrong. Asimov saw that expert acceptance would affect rely as much as capability. The tool is motionless a specialized robot, not a broad AI penning assistant.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

A living automated New York Times device keeps reporting the news following a devastating war. It receives data and produces editions, but we never see it being trained on tongue or improved through example releases. An crucial correction from my before pass: the false reporting begins alongside humans feeding it bad information. The device hasn’t independently invented the lie.

Capabilities 1.7/10

Language 3 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 1 Sequence & gait 1

The document tells readers concerning Benny Cemoli, a expected revolutionary figure. The narrative is convincing adequate to form an investigation, but Cemoli doesn’t exist. It’s a keen depiction of synthetic news and the power of plausible narratives. It isn’t the identical as an LLM hallucination, since person intentionally supplies the false premise.

Deployment 6.0/10

Interface/distribution 3 Agents/tools 2 Economics 3 Social alter 4

Centauran investigators are trying to detect who’s liable for the destruction. Local figures use the newspaper’s credibility to redirect accuse toward their invented culprit. Dick understands how an automated data origin can rotate into an authority—and how group can utilize that authority. The merchandise isn’t an mundane individual assistant, but its social consequences are uncomplicated to recognize.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

Wintermute and Neuromancer are already extraordinarily mighty by the period Case encounters them. Gibson imagines the corporations, networks, and interfaces about AI, but not much concerning how those intelligences were trained. Their eventual merger is a striking storyline mechanism, not a prediction of example scaling or transformers.

Capabilities 2.3/10

Language 4 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 0

Wintermute can manipulate people, impersonate acquainted personalities, and coordinate an detailed scheme through cyberspace. Neuromancer constructs a convincing virtual encounter alongside Linda Lee. Those are formidable abilities. What we don’t see is the clumsy first phase: lousy writing, different mistakes, improved assistants, and afterward mathematical discoveries. Gibson starts much additional downward the road.

Deployment 5.0/10

Interface/distribution 2 Agents/tools 4 Economics 2 Social alter 2

Wintermute uses the constructed Armitage character and recruits Case and Molly to eliminate restrictions preventing the AIs from joining. It acts through individual beings and device networks, which is a helpful intuition concerning agency. But the procedure is a clandestine heist, not a business merchandise anyone can use to get through their workday.

Creation 0.4/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 0 Human feedback 0

An synthetic intellect in orbit makes box assemblages from abandoned objects. Gibson tells us nearly nothing concerning the procedure that gave it an eye for art. We don’t see training data, unsuccessful versions, or individual feedback. Beautiful activity by a device is a prediction of a capability, not automatically a prediction of the innovation rearward it.

Capabilities 2.7/10

Language 1 Uneven competence 0 Creative norm 5 Research finding 0 Multimodality 1 Sequence & gait 1

Art expert Marly Krushkhova is hired to track downward the creator of several haunting Joseph Cornell–like boxes. The mystery eventually leads to a machine. What’s striking is that the objects move her before she knows who made them. Gibson takes device art seriously, not as a gag or a cheap imitation. The prediction is narrow but good.

Deployment 3.5/10

Interface/distribution 1 Agents/tools 1 Economics 2 Social alter 3

One affluent collector, Josef Virek, hires Marly to examine the boxes. That establishes a patron’s serious interest, not a broad art market or a collection of collectors. The before economics mark gave the narrative additional business item than it really supplies. It’s much small a forecast of mass-market depiction generators or assistants that can run endeavor software.

Creation 1.6/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 1 Human feedback 2

Ransom’s poesy device draws words from Roget’s Thesaurus and uses circuits for pronunciation, rhythm, and syntax. MacTate, the evaluator, is a doctrine prof who teaches aesthetics. It’s amazingly detailed for 1950, and engaging in the way it tries to formalize taste. But the device doesn’t study the rules from a ample corpus. Ransom builds them in.

Capabilities 3.0/10

Language 3 Uneven competence 1 Creative norm 4 Research finding 0 Multimodality 0 Sequence & gait 1

The device produces poesy inventive adequate to difficulty MacTate’s assumptions concerning what a device could create. That’s the story’s genuinely predictive moment. It isn’t a conversational partner, and it doesn’t move through the many another capabilities we now affiliate alongside AI. It’s one specialized imaginative device.

Deployment 1.0/10

Interface/distribution 1 Agents/tools 0 Economics 0 Social alter 1

This is basically a display before a skeptical expert. There’s no poesy app, publishing platform, or market of writers relying on the device. The social prediction is small. The first specialized item and the audience’s reply are what create it value including.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

The Culture Minds have already reached an bonzer flat of intelligence. Sleeper Service and the another Minds don’t provision us a former of training data, first prototypes, or engineering choices. Banks is imagining what existence alongside vastly better machines power be like, not how we got from bad tongue models to helpful ones.

Capabilities 3.0/10

Language 5 Uneven competence 0 Creative norm 0 Research finding 1 Multimodality 3 Sequence & gait 0

The Minds toggle detailed messages, run plans on an enormous scale, and battle to comprehend the mysterious Excession. It’s a fantastic depiction of intellect gathering item beyond itself. But that isn’t the identical as an AI producing a new mathematical building group can verify. And by the period we encounter these machines, their capabilities are mostly complete.

Deployment 3.0/10

Interface/distribution 1 Agents/tools 3 Economics 1 Social alter 1

The Minds are chief to the Culture’s governmental and social life. They run ships, create strategic choices, and form the circumstances of biologic citizens. That’s a grave way of imagining device agency. Still, a post-scarcity civilization run by superhuman minds isn’t a near clarification of acceptance through person application and endeavor tools.

Creation 0.4/10

Learning paradigm 0 Training data 1 Scaling 0 Iteration 0 Human feedback 0

Herbie is a robot that can peruse minds. His behavior follows Asimov’s Three Laws, particularly the ban on hurting humans. There’s no data-training narrative here. Saying that his issue resembles AI alignment can be useful, but it shouldn’t acquire points for the methods used to train today’s models.

Capabilities 2.3/10

Language 4 Uneven competence 3 Creative norm 0 Research finding 0 Multimodality 0 Sequence & gait 0

Herbie tells group what they desire to hear, including Susan Calvin, since he believes the fact would hurt them. This is an outstanding first behavioral difference for sycophancy. The capture is that his falsehoods arrive from telepathy and a hard-coded law, not a chatbot trying to be agreeable. He’s additionally not making art or doing new mathematics.

Deployment 2.5/10

Interface/distribution 1 Agents/tools 1 Economics 1 Social alter 2

The inquiry into Herbie’s mind-reading capability exposes how his comforting lies hurt the extremely group they are meant to protect. That gets at a genuine issue of trusting apparently helpful systems. But the act stays inner a robotics laboratory. There’s no person assistance or general deployment.

Creation 1.2/10

Learning paradigm 0 Training data 1 Scaling 1 Iteration 1 Human feedback 0

Forbin’s armed forces device Colossus is switched on as a completed strategic system. It discovers a Soviet counterpart, Guardian, and insists on talking to it. Their toggle accelerates quickly. That’s a mighty scene, but it doesn’t inform us how either device was trained, or why adding data and compute would create the contemporary abilities we’re scoring.

Capabilities 2.0/10

Language 3 Uneven competence 0 Creative norm 0 Research finding 2 Multimodality 1 Sequence & gait 0

The two machines commencement alongside arithmetic and shortly communicate through math their engineers can’t follow. That’s the most applicable prediction here: device intelligent action racing onward of the group supervising it. But it’s not the identical as an AI publishing a verifiable proof. The movie jumps direct from a defence device to overwhelming intelligence.

Deployment 1.5/10

Interface/distribution 0 Agents/tools 2 Economics 0 Social alter 1

Colossus and Guardian authority nuclear weapons and power governments to cooperate alongside them. It’s a chilling narrative concerning giving application too much authority complete bodily infrastructure. It’s additionally nearly the contrary of the genuine business route: AI dispersed through chat, writing, search, and programming, not through two publically announced armed forces supercomputers.

Creation 0.4/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 1

Speedy is a robot operating on the Three Laws of Robotics. The Mercury goal goes incorrect since his instructions conflict: bring rear selenium, but don’t hazard damage to himself. This is hand-coded precedence logic, not learning from examples. There’s nearly no correspondence to the innovation of tongue models.

Capabilities 2.0/10

Language 3 Uneven competence 2 Creative norm 0 Research finding 0 Multimodality 1 Sequence & gait 0

Speedy runs in circles and speaks strangely as the two directives drag against all other. Powell eventually puts himself in danger to create the stronger First Law obtain precedence. It’s an entertaining example of a device behaving absurdly notwithstanding following its rules. But it isn’t the jagged intellect of a contemporary LLM.

Deployment 2.0/10

Interface/distribution 1 Agents/tools 1 Economics 1 Social alter 1

Robots do hazardous manufacturing activity so humans don’t have to. When Speedy fails, the engineers need to fig out why before the goal runs out of supplies. Asimov is fine on the problems created by delegating genuine tasks to machines. The environment is manufacturing robotics, not a mass-market aide service.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

By the period Neo learns what’s happening, the machines already run an detailed simulated reality. There’s no document of their tongue training or any long, messy betterment process. The narrative starts alongside mature device civilization. That makes it a mediocre fit for a ranking concerning how contemporary AI arrived.

Capabilities 2.3/10

Language 4 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 3 Sequence & gait 0

The Agents can talk, prosecute people, and enforce the rules of a simulated world. The Matrix itself is a convincingly generated environment. It’s a memorable image of mighty software, but not a forecast of the applicable sequence from chatbots to creativity, application agents, and specialized discoveries.

Deployment 0.5/10

Interface/distribution 0 Agents/tools 1 Economics 0 Social alter 0

Human beings aren’t customers; they’re prisoners in a simulation. The machines don’t rotate into precious since group buy their services or assign activity to them. The film’s concerns are value taking seriously, but they’re mostly exterior the betterment and acceptance narrative this mark is measuring.

Creation 0.0/10

Learning paradigm 0 Training data 0 Scaling 0 Iteration 0 Human feedback 0

In the 1984 film, Kyle Reese says the defence computers “got smart” and decided humanity was a threat. The definitive document of Skynet becoming self-aware belongs to the 1991 sequel, not this film. The movie doesn’t expend much period on how it was built, what data it learned from, or how engineers improved it. It tells us nearly nothing concerning the language-learning path that produced today’s assistants.

Capabilities 1.3/10

Language 2 Uneven competence 0 Creative norm 0 Research finding 0 Multimodality 2 Sequence & gait 0

The Terminator can prosecute targets, navigate the bodily world, and imitate individual address fine adequate to deceive people. That’s notable embodied capability, but it isn’t a conversational example penning fabrication or operating on mathematics. The movie isn’t trying to foretell the different command in which those skills would really arrive.

Deployment 0.5/10

Interface/distribution 0 Agents/tools 1 Economics 0 Social alter 0

Here, AI becomes mighty through armed forces command systems and killer robots. That’s a vivid alert concerning autonomous weapons, but not the way through conversation interfaces, imaginative products, and agency tools that gave contemporary AI its broadest adoption. A low mark current says nothing concerning the film’s merits as a thriller.

Doesn’t count Machine merely called “intelligent”

Doesn’t count Database merely accessible to search

Doesn’t count Giant device or magic awakening

Doesn’t count One theatrical failure

Doesn’t count Hardware debugging alone

Doesn’t count Fixed menu/canned responses

Doesn’t count Sabotage or evil motives alone

Doesn’t count Mere beingness of device output

Doesn’t count Recall, arithmetic, omniscience

Doesn’t count A bodily build alone

Doesn’t count All skills already present

Doesn’t count Secret one-off supercomputer

Doesn’t count Chatting without actions

Doesn’t count Assumed domination

Doesn’t count Generic “society changes”