Cass Cain is ready to HERO you

Regarding Deku and his classmates’ presence on the battlefield, one of the things I’ve talked about in the past is “spotting a series its premise.”  That is, if you’re going to get anywhere in a story, you have to allow it its basic scenario or you might as well not bother with it at all.  In BNHA, that basic scenario is “Hero Academy.”  The idea of a school for heroes is presented as something cool and positive, and if you can’t deal with that at all without starting to talk about child soldiers and police states, then BNHA was never going to be the series for you.

BNHA plays its premise of a school for Heroes with fantastical optimism, so it suddenly diving for a “realistic” ending, one grounded in a deeply pessimistic view about redemption and rehabilitation, is an active betrayal of its own tone and themes because its author lacks the courage to follow through on what his main character wants to believe.  By “wants” here, I don’t mean that Deku has some kind of independent life and desires to believe something his creator won’t let him act on.  Rather, I mean that Deku’s in-universe characterization and actions should be harmonious with Deku as a narrative device and embodiment of the story’s overall message.

Based on Deku’s presentation through the first war arc, he should have stuck to his guns about saving Shigaraki!  He should have been willing to defend Shigaraki against All For One and Hero Society and Shigaraki’s own fatalism!  Instead, though, he cowered in the face of all of them.  He didn’t just fail to save Shigaraki; he actively killed Shigaraki.  The tension (which had always existed in Deku) between his admiration of the Hero System and his innate, overwhelming drive to save should have resolved in favor of saving, not in favor of upholding a status quo he knows to be riddled with problems!

And that being the case, of course he should have been on the front lines.  He’s the main character and it was the last battle!  Where else should the main character have been?  The trouble with Deku being the focal point in the final battle isn’t that it turns him into a child soldier; he was always that because it’s inherent to the premise of BNHA and the structure of young adult adventure stories more broadly.  The trouble is that the story fails to convincingly justify his presence under its own previously established standards.

Like, if you look at how U.A. is stated to normally approach its classes, it’s obvious that the situation with Class 1-A is the result of unprecedented challenges!  I’m totally willing to roll with that as part of spotting the series its premise!  It’s when the series stops being able to justify 1-A’s role in events that my acceptance of the scenario starts breaking down.

Let me go over that in some more detail:

Internships: Per the story, internships are heavily chaperoned and focused on light work, really just dipping a student’s toes into what a normal day in the life of a Hero looks like, including all the non-combat procedural stuff and the social angles of endorsement deals and the like.  Even a hardass like Gran Torino states that his intention is to take Deku to someplace where “minor crimes” are common, not to plunge him right into the deep end.  Indeed, a fight breaks out, he firmly orders Deku to stay on the train.  When Deku pursues, the first Hero who notices him tells him to follow police instructions and get to safety.  It’s obvious that the focus is on a safe first-time experience—it’s even observed later on that the internships basically treated the kids as “guests,” and they weren’t allowed to do anything dangerous.

Work Studies: These aren’t even supposed to be offered to first-years.  The only reason they are is that Class 1-A has been repeatedly targeted and thus, amidst a very divided staff opinion, U.A. made the choice to offer them to the first-years to better prepare them for the increasing danger and uncertainty that is both targeting them personally and becoming more prominent in a post-All Might world.  U.A. personally vetted agencies offering work studies before approving them.  These are more dangerous, but still strictly overseen, carried out with as much caution as can be managed for a job that inherently carries with it life-or-death risk.

So far so spotting-the-series-its-premise.  The problem starts kicking in harder once the war arc rolls around.  There are hundreds and hundreds of active Heroes in the country; I simply do not believe that the planners of the raid on the PLF needed e.g. Juzo’s earth-molding when they could have tapped Pixie-Bob for hers, or Tokoyami’s darkness-boosted strength when they could have just used Midnight to flood the staircase Re-Destro was using with sleeping gas.

However, there’s still a bit of an acceptable handwave in that we have started to see the corruption and desperation of the HPSC—a number of the adult characters comment on it being kind of Weird and Sus that the HPSC is ramping up combat training for high schoolers, especially first-year high schoolers.  The anime adaptation goes out of its way to note that the students’ involvement in the war was hidden from the media, further emphasizing that the HPSC’s actions are intended to be read as over-the-line and suspicious in ways that the internships and work studies were not.

But that all goes completely off the rails for the second war.  The story wants us to believe that all the HPSC’s corruption was neatly dealt with by Clone!Re-Destro murdering the President (as if she wasn’t succeeded first by a man who also took part in the child assassin program and later by one of the actual child assassins).  Suddenly the kids of 1-A are cornerstones of the battle plans, with the only feeble handwaves given being that the first war reduced the number of Heroes drastically (but somehow didn’t impact any of the Heroes we’re actually supposed to care about beyond Midnight) and that the kids have extended experience with fighting the League (equally absurd; if it was just the experience, they should have tapped the kids for their knowledge during the planning stage, not the actual battle).

It’s obvious by this point that the only reason the endgame is rotating around the main characters is because they are the main characters, not because the admirable adult Heroes the story tells us we’re supposed to be rooting for have actual justification to put those kids in that position.

Now, I talked earlier about treating Deku as a character because he is one, so what’s the problem with accepting his presence because of his role rather than trying to examine the realism of the situation?  Well, the problem is still the same: the story is trying to sell us one thing (realistic explanations for the kids’ presence that respect the prior worldbuilding and don’t undermine the good guy chops of their whole side) while providing us with something else (transparent contrivance that shrinks the scope of the world while doing nothing but undermining their good guy chops).

If the story had Deku stick to his principles on saving Shigaraki, to the point that he was willing to depart from the entire professional structure that was telling him he couldn’t, then it would be perfectly reasonable for him to have to fight Shigaraki one-on-one!  Indeed, he have had to, and worse, he might have had to fight off his own erstwhile allies too.  Conversely, if he’d stood up to the planners of the war and convinced them to believe in his intentions regarding Shigaraki, the same way he did the prior bearers of OFA,[2] then it would also make sense for him to be the central point of the battleplan.

2: Give or take the shamelessly easy out Horikoshi took by having Yoichi convince Kudou and Bruce rather than making Deku find a way to do it.

But Deku doesn’t do any of that.  He passively lets himself be steered into a confrontation with a guy he wants to help but refuses to defend against others’ violence or to engage with honestly.  The battle arena for Shigaraki is literally called, in-universe, the Flying Coffin; all the planners wanted to do at that point was kill Shigaraki.  Deku never even whispered a protest, never so much as asked for a chance to talk to Shigaraki before firing up the electrical field.  So not only do I not believe the Heroes' justification for putting Deku here, I don't believe Deku as-established by the story should have let himself be put here. The only Watsonian explanation is that Deku is a much lesser Hero than I thought he would be, and yet the story is still trying to convince me that he's this amazing Hero whose “mad” “drive to save” “eclipses all common understanding.”

You, anon, can’t tell me I have to credit Deku for doing the bare minimum when the story is telling me he changed the world.  You can't tell me I have to see him as someone who might start realizing his society is fucked up and perhaps will start trying to do something about it when to the bitter end all he could talk about was about how he planned to Bring It All Back, with all talk about doing things better falling to characters as minor as Mirio or Jirou.  You can't tell me it makes sense that he failed when the story is frantically trying to spin his failure into a feel-good success.

Deku being sad about killing Shigaraki for approximately a day and a half does not negate the story depicting Deku's defeat of Shigaraki as him literally punching the sunshine back into a stormy sky.

Deku told me that this was the story of how he became a great Hero, he told me that for him being a Hero meant saving people, and then he failed to save the person his whole arc was leading him to save.  And then he had the gall to tell me that the point of the story wasn’t about him becoming a great Hero after all, it was actually about “reaching out to help,” regardless of whether that reaching out is actually successful or not!

Timnit Gebru was fired from Google in December 2020 for refusing to retract a research paper, and every single warning that paper made about large language models has now happened at a scale the industry spent 4 years trying to make people forget about.

Her name is Timnit Gebru.

She co-led the Ethical AI team at Google. She co-wrote a paper called "On the Dangers of Stochastic Parrots" with Emily Bender at the University of Washington and two other researchers. The paper was 14 pages long. It was submitted to a top AI ethics conference. And it was the reason Google decided that one of the most senior Black women in AI research could no longer work there.

The story Google told publicly was that she resigned. The story she told, confirmed by 2,695 of her colleagues in an open letter, was that she was fired by email while on vacation because she refused to either retract the paper or remove her name from it.

The paper had not even been published yet.

Here is what she actually wrote, and why every prediction inside it has now come true.

The first warning was about scale itself. Bender and Gebru argued that training ever-larger models on ever-larger scrapes of the internet would produce systems that appeared fluent but had no actual understanding of language. They called these systems stochastic parrots because they would repeat patterns from training data with statistical confidence and zero comprehension. The paper predicted that this apparent intelligence would fool both users and developers into trusting outputs that were structurally incapable of being reliable.

This was 2020. GPT-3 had just come out. The paper predicted the hallucination problem before anyone had a word for it.

The second warning was about bias amplification. The paper documented in detail that internet-scale training data contains systematic overrepresentation of dominant viewpoints and underrepresentation of marginalized ones. The models would not just absorb this bias. They would amplify it, because the optimization process rewards confident outputs, and confidence in language patterns tracks frequency in the training set.

The prediction was that hiring tools built on these models would discriminate against women. That healthcare triage tools would underperform on Black patients. That loan approval systems would entrench inequality while presenting their decisions as neutral algorithmic judgment.

Every one of those things has now been documented in deployment.

Amazon's hiring algorithm penalized resumes that contained the word "women" in any context. Healthcare risk scoring algorithms used by major US hospitals were found to systematically underestimate the medical needs of Black patients. Apple Card's credit algorithm gave wives credit lines 10x lower than their husbands for the same financial profile.

The third warning was about environmental cost. The paper calculated that training a single large language model produced emissions equivalent to the lifetime output of 5 cars. The prediction was that the race to scale would create an environmental footprint that would eventually rival entire industries.

In 2024, Google's emissions were up 48% from 2019, and the company explicitly blamed AI infrastructure. Microsoft's were up 29%, same reason. Both companies have now quietly abandoned the climate commitments they were publicly celebrating the year Gebru was fired.

The fourth warning was about documentation. The paper argued that the training datasets being assembled were too large for anyone to actually audit. Nobody at Google, OpenAI, Meta, or any other lab could tell you with confidence what was in the data their models were trained on. This was not a temporary problem to be solved later. It was a permanent feature of the approach.

In 2023, researchers discovered that the LAION-5B dataset, used to train Stable Diffusion and other major image models, contained thousands of images of child sexual abuse material. The companies that had trained on the dataset had no way of knowing. The paper predicted that category of failure 3 years before it was found.

The fifth warning was the one Google cared about most.

Bender and Gebru argued that the deployment of these systems would centralize linguistic and cultural power in the hands of the small number of companies that could afford to train them. The internet would become a place where the dominant voice was a statistical average of dominant voices, presented as a neutral assistant. Languages underrepresented in the training data would degrade over time as more web content was generated by these systems and fed back into the next training run.

This is now happening in real time. A 2024 study found that 57% of new web content in English is AI-generated or AI-assisted. Researchers studying low-resource languages have documented active degradation in translation quality, because the synthetic content fed back into training is itself worse in those languages.

The paper Google fired her for predicted the model collapse problem before model collapse had a name.

The mechanism behind why this all happened is the part of her work that nobody quotes.

Gebru's argument was not that AI is dangerous in some abstract sci-fi sense. Her argument was that AI is dangerous in a very specific structural sense. The technology was being built by a small group of researchers who shared similar backgrounds, worked at similar companies, and were rewarded for shipping products faster than competitors. The incentive structure made it impossible for safety, ethics, and bias concerns to slow anything down. Anyone inside the system who raised those concerns was either ignored, sidelined, or removed.

She was making that argument from inside Google.

Then Google proved her right by removing her.

The team Google had built to make sure their AI was safe was dismantled in 90 days because they did the job they had been hired to do. Margaret Mitchell, the other co-lead of the Ethical AI team, was fired two months after Gebru for searching through her own emails for evidence of how Gebru had been treated.

Gebru did not stop. She founded DAIR, the Distributed AI Research Institute, in 2021. The mission is to do AI research outside the control of the companies that have a financial interest in not hearing the answers.

Every prediction in the Stochastic Parrots paper has now been validated by deployment. Hallucinations are an industry-wide problem the largest labs cannot solve. Bias amplification has been documented in hiring, healthcare, lending, and criminal justice. Environmental costs are larger than entire small countries. Training data audits remain impossible. Model collapse is an active research crisis at every major lab.

The question worth sitting with is the one almost no one in the industry will say out loud.

Every researcher with the technical credibility to call out these problems watched what happened to her in December 2020 and made a calculation about their own career. The number of people willing to speak publicly about safety and ethics issues inside the major AI labs collapsed after that firing and has not recovered.

The researcher Google fired for warning about exactly what is now happening was right.

The company that fired her is now the second-largest deployer of the technology she warned about.