Star Trek: The Original Series

"The Ultimate Computer"


Air date: 3/8/1968
Teleplay by D.C. Fontana
Story by Laurence N. Wolfe Directed by John Meredyth Lucas

Review by Jamahl Epsicokhan

Starfleet informs Kirk that the Enterprise is to serve as test subject for the new M-5, a groundbreaking advancement in computer technology, designed to make command decisions faster than captains and reduce the number of people required to run a starship. An astute allegory for contemporary automation at the expense of "the little guy," this episode's first few acts are superb, as Kirk finds himself debating whether he's selfish for wanting to keep his job at the expense of technological progress, or if it's a matter of actual danger or principle.

A wonderfully acerbic debate between Spock and McCoy about the role of computers is also well conceived, ending in Spock's well-put notion to Kirk, "...but I have no desire to serve under them." Following the M-5's initial success, the scene where another captain calls Kirk "Captain Dunsel" is the episode's best-played and simultaneously funny and painful moment. (In a word, ouch.)

Once M-5 runs out of control and hijacks the Enterprise—resisting attempts to be shut down in acts of self-preservation (including murder and eventually full-fledged attacks on other Federation starships), the episode turns to an frightening analysis of M-5's creator, Dr. Richard Daystrom (William Marshall), a man obsessed with outdoing his prior successes, who has created a monster that he has come to regard as a child. Though it pushes a little hard toward the end (Shatner and Marshall going a bit overboard), the story is a compelling one.

Previous episode: The Omega Glory
Next episode: Bread and Circuses

◄ Season Index

20 comments on this review

Thu, May 16, 2013, 12:56am (UTC -5)
This is an excellent episode, but its strong characterization of Kirk falls down with an ending that finds him grinning and chuckling at Spock and McCoy's verbal jabs as the music takes us out on an upbeat everything's-peachy-again tone.

The problem is that the entire crew of the Excalibur has just been murdered, along with a good chunk of the crew of the Lexington. Some 500 men and women dead, a horrific tragedy that's made even worse by the fact that the Enterprise was the instrument of their destruction. There's no way the bridge crew ought to look this happy in the closing moments, and Kirk, knowing that the ship he so loves was used to do such a terrible thing, ought to be truly anguished.

If this had been a first-season episode it probably would have ended on a somber note, but the second season got considerably lighter and "The Ultimate Computer" was only one of a number of eps that year to end with inappropriate humor.
Fri, Jan 10, 2014, 1:59am (UTC -5)
Brundledan expresses my thoughts and then some. The episode itself fails to take it's own lofty premise seriously. Other things that I disliked were:

- Kirk's apparent crankiness toward Daystrom even before anything went wrong, and after some lengthy and deep on-screen self-reflection.
- The implausibility of a major military organization like Starfleet allowing this test to be carried without proper testing, training and precautions, as well as the possibility of such a flawed computer as the M5 of ever being granted a test run.
-When Commodore Wesley assumes that Kirk is responsible for Enterprise's attack on the Excalibur, even though he originally browbeats Kirk throughout the prelude to this mission ... and then suddenly comes to his senses and calls off the subsequent attack, thereby killing off alot of the tension and drama that had been built up for a climactic scene.

I get that some of these elements were put in place to set up the story as a drama of one man's (Daystrom) obsession with his creation, but this was an element that seemed to coalesce rather late within the story, and lacked relatability.
Fri, Jan 10, 2014, 9:17am (UTC -5)
@Alex: I think you can chalk up Starfleet letting Daystrom test the M5 and not checking its flaws by chalking it up to reputation. Also, Kirk understood almost immediately that the M5 could cost him his job, so being cranky toward Daystrom made some sense.

But you're absolutely right about Wesley. His first scene in the transporter room is off-kilter ("Hey old friend -- congrats on losing your job!"). Then (as you noted) it's weird that he would blame Kirk for the attack later but also think enough of him to hold off on firing on the Enterprise at the end.

But the entire episode is oddly characterized. Even for Shatner, Kirk is over the top in this one. His scream of "DAYSTROM!" near the end was really strange. Nimoy does his normal nice job, but even Kelley seems like he overacted ("That thing just destroyed an ore freighter!!!").
Thu, Mar 20, 2014, 3:04pm (UTC -5)
This is an odd episode. There's a lot of really strange characterization -- like the opening scene with Commodore Wesley. He acts quite odd to a friend who, essentially, is losing his job.

Also, the Kirk/Daystrom stuff at the end of the episode is just over the top.
jonn walsh
Sun, Jun 7, 2015, 7:37am (UTC -5)
A favorite of mine but there is a plot aspect that nearly kills it for me.
This is not Kirk's job. The Federation isn't merely a military organization...We're out there exploring and introducing our flavor of diplomacy and friendship to all the other species 'out there' by choice. We're out there because WE WANT TO BE. Starfleet personnel, Admirals, Captains and indeed, you would believe their crew as well, are living their dream. M-5 would simply be another tool at our heroes disposal.
Kirk, et al, are explorers and ambassadors in the final frontier performing interpersonal functions with other species that a computer could not begin to assume. No one's 'job' is in danger. It's an artificially created and inflated plot point.
Wed, Aug 19, 2015, 4:26pm (UTC -5)
I liked the episode overall but the whole Captain Dunsel thing was very offputting. It was quite inappropriate and disrespectful of a star fleet commodore to call another captain something like that.
Thu, Aug 27, 2015, 11:02pm (UTC -5)

My name is Captain Dunsel.

I'm sorry my command of the Enterprise did not go well.

I've been demoted to ship's junior cook, under some dude named Neelix.
Jason R.
Wed, Dec 2, 2015, 2:40pm (UTC -5)
The whole episode is a big cheat, and epitomizes Roddenberry's enduring failure to take AI seriously. It's a cheat, because the only way Roddenberry finds to nullify the AI as a threat to human commanders is to arbitrarily make it homicidal and insane.

Roddenberry's approach to AI, as exemplified in this episode, always struck me as cowardly and backward. There is a principled argument to be made for the need for human beings, even in the face of increasingly capable AI, but this is not it.

It was not until STNG - The Measure of a Man that the series really began to take AI seriously, if only a little, and I'm assuming by that point Roddenberry was on the way out in terms of his influence on the show.
Wed, Dec 2, 2015, 7:51pm (UTC -5)
Well, Jason R., if the story was intended to demonstrate the superiority of the human factor versus machine intelligence, then that thesis is undercut by having the machine's flaw being traits copied from its human creator. So maybe the story is more complex than simply "human 1, AI 0." Or it is that simple and they botched it.
Jason R.
Thu, Dec 3, 2015, 10:09am (UTC -5)
Yeah Grumpy, actually I thought about that irony after I finished this comment. While that is a clever interpretation, I just think it's giving too much credit to the writing. The plot device that makes the computer insane is really incidental.

Roddenberry doesn't ever seem to be willing to confront the idea that a computer might very well be superior in some respects to humans. His portrayal of AI in general has been lackluster. Data was the first serious attempt to do so, and even then Data's central premise was a desire to be more human.

I again don't fault Roddenberry for believing humans to be inherently superior to AI - I just fault him for lacking the imagination to present this thesis in a halfway intelligent fashion.
Peter G.
Fri, Apr 1, 2016, 3:51pm (UTC -5)
@ Jason R.

I think the point of this episode is pretty clearly that an AI's effectiveness is limited by the limits of its creator. If humans are flawed that means an AI will be flawed as well; its weakness will be a reflection of human weakness, except you can't reason with the machine.

While I agree with you broadly that Trek in general avoids the subject of AI, within the confines of this specific episode I think it approaches its subject matter very well. What Roddenberry thought at the time was likely quite correct at the time - which is that limitations in computer programmers would create severe limitations in computers. I doubt very much he foresaw the possibility of AI as a self-evolving system, even to the point we have now where a recursive program can teach itself activities such as playing Go such that it doesn't have to be completely programmed from the get-go. And he surely didn't think of AI in terms of quantum computing or bio-neural circuitry that could do computations at a godlike speed.

I personally prefer to attribute this lack of vision to the limitations on AI as they existed in the 60's, and to the fact that Trek is predominantly supposed to be a person-oriented show rather than hard sci-fi that deeply explores what various tech can do. It is for this reason that I prefer physics novelties such as we see in TNG to be backdrops to episodes rather than the central plot point to be solved. "Cause and Effect" is a great example of what I like, because it employs a novel physics idea as a backdrop but the real action is a character-driven mystery story.
Mon, Apr 18, 2016, 12:38pm (UTC -5)
@John Walsh The way they talk about it in the episode, it seems clear that there would still be crew, they simply do not need captains and certain personel.
Thu, Feb 23, 2017, 9:25pm (UTC -5)
Sigh, Kirk outwitting a computer... again.

I mean, I get what they were trying to do with this episode, and I understand the uncertainty of how these newfangled computers would fit into society back in the 60s and all, and I realize it was a hot topic in sci-fi, both profound and silly, but man, there's nothing that makes an episode look more dated. All these old shows and stories assumed you would feed a few pieces of information to a computer, and it would make surprising connections and leaps of logic that would shock and amaze people. Hey, maybe that will still happen in the future, who knows? But our modern computers seem to be on par with the normal Enterprise computer these days, and this idea is still well beyond our comprehension. So it just seems bizarre that all these SF shows completely missed all the ways computers would actually impact our lives and jumped straight into these superbrain stories. And, just like all those other SF stories, the ultimate computer ends up being evil.

And that's what really bugs me, makes me think this episode is not a true classic. The computer just goes straight to evil. The character arc or struggle or theme of the episode was Kirk worrying about being replaced, being obsolete. That's a fair story to consider, so the antagonist of the plot (the computer) should be one that complements and reinforces this struggle. But instead, it just acts as a straw man. The resolution should be that Kirk isn't obsolete because he has some unique quality that the computer doesn't have. Think about, for example, the Corbomite Maneuver, where only a bluff would work to save the ship. Or all the old Kirk speeches to get the antagonists to change their mind. A proper resolution would show that Kirk had something above the computer, like the battle of wits between him and Khan. Instead, he doesn't have to show why he deserves to be a captain when it became clear that the computer is crazy evil. Just shut off the computer, abandon the project, and fly off into the sunset, and no more self examination of Kirk. Is that really what we wanted?

Thus, a potentially interesting idea went to waste. Too bad.
Peter G.
Fri, Feb 24, 2017, 12:07am (UTC -5)
@ Skeptical,

You may argue that the episode didn't create the greatest case for man over computers, but I think you would be wrong to suggest that it failed to create a case altogether.

The point of the episode isn't just that Kirk is smarter than the computer, or that no computer can match a human in creativity. That may or may not be true, but it isn't exactly the point. TOS always has as a running theme that logic and computation alone isn't enough to make a great person or a great society; this is reflected repeatedly in the Kirk-Spock-Bones trio. Kirk isn't just logic, but is logic coupled with humanity and compassion (Bones + Spock = Kirk). The fact that the episode (as usual) ends with the computer being 'outsmarted' is a tidy way to wrap things up, and I agree that it's a weaker ending than it should have had. But the wrap-up isn't really the point as I see it. The point is that a machine will follow its logic to the end and have any fallback position grounded in compassion, sympathy, or feeling. It's sort of like a psychopath, if you will, in that it will not have internal mechanisms to stop it doing bad things if they seem best.

Now, it's true that if the programming is good then the output should be ok too, and likewise if there is a bug (a la Skynet) things will go pear shaped and the computer will not be able to be reasoned with past that point. But more to the point, the Trek theme is TOS is that advancing humanity isn't about technology or capabilities, but primarily about advancing values and how we treat each other. This is an area in which the inclination to push capability will not only be a sidetrack to advancing humanity but will in fact hinder it if pursued incorrectly. Take, for instance, the eugenic wars, where in an effort to 'advance humanity' in capability a monster was created instead. Likewise here, where a captain more sophisticated than a human is created to obsolete humans, just as Khan wished to obsolete homo inferior. The danger outlined in "The Ultimate Computer" is along these lines, and although it didn't fully realize the treatment of this issue I do think it's in there and is still pertinent to this day; maybe more so than even it was in the 60's, when human obsolescence was still science fiction.
William B
Fri, Feb 24, 2017, 2:55am (UTC -5)
Interestingly, I was going to make something of the opposite point to Peter, though in a way that is not inconsistent. The episode actually suggests that the M-5's value is not just because it can do normale computer things, but because it goes beyond usual computing into the domain of people -- creative thinking and all that. Specifically, this is because Day Strom programmed it with his own memory engrams. When the computer goes haywire, it is because it has inherited Daystrom's flaws as well as his strengths. I tend to see the message of this particular element as that computers are still created and programmed by people, and so will always be limited by the people who made them. The computer's apparent usefulness was that it could match human genius without flaws, but that was wrong, and the reveal that there is no machine utopia allows Kirk's Imperfect humanity to be back in command. There is a similar story in TNG where Data and Lore "inherit" some of Soong's flaws, though this is much more pronounced in Lore and Data was deliberately created to be aware of his limitations and to want to coexist with rather than dominate humans.

To build off Peter's point, Daystrom going insane may be a way of showing that the danger of thinking that machines can supplant, rather than supplement, humans is that the humans who subscribe to this may develop their own machine-like flaws. Daystrom's inability to think of the universe in terms besides efficiency and the attainment of his goals (and his inability to conceive of his own worth) make him kind of computer-like, as does his social isolation. This ends up enhancing his human flaws, which again seems to result from hanging his identity on a dream of escaping from human flaws entirely. I think the end can be both that Daystrom has made the mistake of thinking like a machine, and that M-5 is dangerous because it "thinks" like a person, though it is maybe a bit complicated.
Fri, Feb 24, 2017, 7:47pm (UTC -5)
Well, I still stand by my statement, that the implementation of the episode to complement the theme was not done well at all.

William, you seem to state that the theme shows by comparing Daystrom erraticness to the M-5's going cuckoo. And yes, perhaps that is the reason the M-5 did poorly, since it was his brainwaves that he used to create the M-5. And therefore, you say, the theme is that the people who want to supplant humans have their own problems that preclude them from being the best judge of humans. Well, ok. But still... well, it's obvious the theme the episode wanted to show was that humans are not going to be obsolete by this computer, what with the whole "dunsel" bit. And if William's interpretation is true, then I, like Garak and the Boy Who Cried Wolf, see a different moral. If the importance is to show the connectivity between Dyson and the M-5, then the moral of the story isn't that computers are inferior, but rather that better humans should be used as the template for computers.

After all, if the fault of the M-5 is just that Dyson was erratic, why not try the M-6 with Kirk's brain? Will that computer be perfect enough to replace human captains? I don't think the episode answered that. Which is why it's a bit of a straw man story - it's not really Kirk vs The Ultimate Computer. It's Kirk vs the Insane Computer. And that's not a fair comparison.

Peter, I don't really disagree with what you say. But I just feel a bit more strongly on the fact that it's weak than you. Yes, computers will follow their logic to the bitter end, which can seem horrifying. And yes, it does mean that there should be some human oversight. Which honestly should have been obvious, but of course they didn't show it. Naturally our superintelligent future means people will test a complex new computer by giving it complete control of a freaking battleship that has enough firepower to exterminate a planet, and make the only kill switch an electronic one that the computer can hack.Perhaps it should be Starfleet command that should be seen as insane...

But I digress. My problem is that I, Robot came out in 1950. The Three Laws were first introduced in 1942. While Trek may have been blazing a trail for television sci-fi, this episode feels 25 years behind the times when it comes to sci-fi in general. There should have been safeguards put in place on that computer. There should have been better logic programmed into it. But apparently, Dyson didn't think of it. And apparently, Starfleet didn't demand it before thinking about putting it in one of their ships. It just wasn't very intelligent plotting, and so it's tough for me to care about the theme when it relies on dumb plotting.

(With that said, I will point out that this episode came out about a month or so before 2001, so it's not the only visual medium showering murderous AI. But HAL is a lot more memorable, so I'll let that one slide...)
William B
Sat, Feb 25, 2017, 5:18am (UTC -5)
Skeptical, it's a good point that the M-6 could be imprinted with Kirk's brain. In my interpretation, the episode erred by having Daystrom go bonkers at the end, because I don't think the point was even that Daystrom was a particularly crazy or bad individual, so much that any computer created by humans (let's narrow the focus from aliens here, this is TOS and pretty human-centric) will inherit human flaws. The issue then is lack of balance. No individual human would be capable of running the Enterprise not just because of the physical or even computational demands, but because humans need constant checks and balances to keep from losing perspective. Kirk is in command, but he has Spock and Bones to constantly play off, and Kirk listens to them. But even if it weren't for that, Kirk has humility not to expect that he can run everything by himself -- or, indeed, the humility to recognize he's not perfect. Actually since Kirk sometimes has mild megalomanic traits, kept in check largely by his close attachment to Spock and McCoy, an M-6 designed on Kirk would also run into the same problems. The delusion is not that the M-5 is capable of running the ship's systems, but that it should and that its "judgment" will remain superior to humans', when it is still based on humans and so will likely not be a magic way of evading well-known human flaws. I think this is part of the point in 2001, as well -- HAL is a tool crafted by humans, and so his programming is still susceptible to "human error," just at a different point and level than human mistakes. Or, rather, HAL works perfectly according to the code as designed by its/his human programmers, and the underlying flaws in their thinking only become exposed once it runs its course, similar to (say) the underlying logic of the doomsday machine system (including both the tech circuitry and also the loyal soldiers following orders) in Dr. Strangelove.

I do see what you mean that it's a strawman because Kirk doesn't actually face The Ultimate Computer. But...I think the episode's point is that there *is* no "Ultimate Computer," or at least it's far further away than people think. If we define the Ultimate Computer as a computer capable of running a starship *technically*, then Kirk could outthink it with lateral thinking as is the case with most of the computers he faces; if the Ultimate Computer is a computer capable of human-style lateral thinking and creativity, as seems to be the case here, then it inherits human flaws along the way and so it is necessary to install the usual checks and balances, which really comes down to wanting a human making the final shots anyway. That Kirk outsmarts the computer in the traditional way here is, I agree, another flaw in the episode -- this computer should be smart enough not to fall for it, or else it *is* just another Nomad or whatever.

The other element, which the episode does talk about, and which I think would be better to look at squarely, is the question of whether computers running things, even if they could be entirely trusted, would be whether human dignity would be removed/ruined by giving power to the machine. And I think most stories still use the idea that computer-run societies will end up being some kind of dystopia to avoid the issue of whether a fully pleasant computer-run world would really be so bad. I still think that the dystopia argument has value because I think that there are lots of reasons to suspect that any system designed by humans will eventually run into human-like problems, but, still, it is hypothetically possible that this is not the case, and then there is still an issue of whether humans should avoid over-reliance on machines for their decision-making, even if those machines are genuinely able to make those decisions better. That's what this episode seems to be about for a time, and I value what it "turns out" to be about...but, yeah, I would also like to see that other story.
Peter G.
Sat, Feb 25, 2017, 8:36am (UTC -5)
@ William,

I wonder whether Daystrom going insane might be intended to mean something more than merely that the machine had a faulty programmer. One of the classic sci-fi elements to an AI dystopia is not that the machines fail, but that they entirely succeed in fulfilling their role. What happens is that instead of machines helping man to achieve his dreams instead they serve as an excuse to stop pursuing them altogether. Instead of helping man to think, they give him an excuse to stop thinking and to turn over his free will and volition to them. From the start I think we get the impression that Daystrom is not only excited about the technology itself, but seems to actually be excited at the prospect of humans being replaced by computers; it's almost a self-destructive fantasy coming to life. As he goes mad towards the end, almost in tandem with the AI, my sense is that this might mean not that he was always flawed, but rather that he had by this time placed all of his hopes into the AI and was dependent on it. When it began to fail he began to fail. We don't know his backstory here and can only guess, but what if he had already been using AI to help guide him? What if the computer itself had assisted his research and maybe even given him the idea to put it in command of a starship? The idea that he had become a servant to a machine could indeed make him become unhinged. Of course this is my own imagining, but broadly speaking I think the sci-fi world was already becoming acquainted with the notion that letting machines take over out thinking for us not only poses a danger due to the machines themselves, but also in allowing us to become dependent on them for everything.

As a complete aside, I'm not sure that the correct interpretation of 2001: A Space Odyssey is that HAL malfunctioned. True, that's the prevailing understanding, but my suspicion, especially knowing how Kubrick thought, is that HAL was programmed to deliberately turn on the crew so that it could contact the aliens by itself and report directly to whomever programmed it, without the crew blabbing.
William B
Sat, Feb 25, 2017, 10:16pm (UTC -5)
@Peter, right, I mean, my point wasn't really supposed to be that Daystrom himself is *particularly* unhinged and always has been. Rather, Daystrom thinks that he's a good model, and the reason is simple enough -- Daystrom is also self-evidently a genius. And under normal circumstances, he would be a good example of what is good in humanity: he's brilliant, creative, altruistic, working toward the betterment of the species. His flaw turns out to be monomania; his obsession with prioritizing the M-5 above all else ended up spilling over into the M-5 prioritizing...itself over all else. But I think that other people have different flaws, which when wedded to an Ultimate Computer-style starship which is expected to fulfill the function of dozens of humans would also be disastrous. I think Daystrom's breakdown suggests both that, as you indicate, he had put too much of his hope in machines, and also that he also was overloaded. We learn that he succeeded early in life, was seen as a whiz kid (something of a human computer) and then has spent the rest of his life trying to live up to those expectations, sort of like Stubbs tells Wesley in "Evolution"; while Daystrom has an inflated ego, it's not simply arrogance but some fundamental lack of conception of his worth outside his success. This overloading is similar, maybe, to M-5's overloading, but it also fits in well with the idea of a person desperately seeking a way to do away with human failings. I'll have to think about it.

In some ways, there is also a parallel between Daystrom and Kirk -- because Kirk also may in fact need to feel useful even if, as he acknowledges at one point, he is *not* needed as captain anymore. Daystrom mostly seems to want to make everyone else obsolete, and there may be some latent sense of revenge on Federation society in it -- he wants to make everyone feel like he felt, after his own tech made *him* obsolete, to the point where his only possible use to society seems to be an apparently unattainable goal. Kirk's ability to question his motives seems to be the thing that sets him apart from Daystrom at this moment -- but this is by no means an indication that Daystrom is congenitally a madman, so much as that extreme fame and adulation followed by inability to meet one's lofty standard create perverse incentives and take a big psychological toll. In fact, maybe that's the trick -- Daystrom, whose own invention put *him* out of work, is the proof of the long-term psychological damage of replacing a person with a machine entirely. Daystrom's desire to have an even better machine seal his legacy by replacing all of humanity is not only self-destructive in the abstract, it's specifically almost a kind of Stockholm Syndrome, repeating-of-trauma -- Daystrom's sense of worth has eroded since his first big breakthrough. (It reminds me of the classic image of a gambler who wins big on his first time out, and then develops a strong addiction because that rush/depression pattern is absolutely set early on, though I do think we are meant to see Daystrom as a genius rather than having succeeded by accident; very few people have one moment of humanity-changing brilliance, let alone multiple ones.)

Good point about HAL. I tend to think that even if he wasn't specifically programmed to kill the humans, he didn't particularly "malfunction," in that he was still following a logical course. The consequences of humans mucking up contact with alien life forms are too great to ignore, and it is logical from a certain perspective to eliminate potential sources of error and to maintain total control in what could be a major turning point in human history. This would make sense even if HAL was entirely programmed to put the mission (and the ultimate good of humanity) as a top priority.
Thu, Mar 16, 2017, 12:04pm (UTC -5)
Why are commodores in star trek always such major dicks? I really enjoyed the scene between McCoy and Kirk where Kirk feels at odds with his ship. And this is my main gripe with Star Trek Continues - the fan made show...when they introduced the counselor they eliminated the need to have any meaningful scenes with McCoy in that particular show - but that's just my opinion.

Submit a comment

Notify me about new comments on this page
Hide my e-mail on my post

◄ Season Index

▲Top of Page | Menu | Copyright © 1994-2017 Jamahl Epsicokhan. All rights reserved. Unauthorized duplication or distribution of any content is prohibited. This site is an independent publication and is not affiliated with or authorized by any entity or company referenced herein. See site policies.