Thomas Edison had staked everything on direct current.
When Nikola Tesla appeared to have succeeded with a system based on alternating current, Edison was furious — and he did not respond by improving DC. He set out to ruin Tesla's reputation by making the public believe "that the AC system was inherently unsafe, and Tesla irresponsible in promoting it."1
The campaign had two stages and both are worth stating plainly, because Greene states them plainly.
He captured household pets and electrocuted them to death with AC current.1
When that wasn't enough, in 1890 he got New York State prison authorities to organise the world's first execution by electrocution, using an AC current.1
Greene's verdict on the outcome:
Although, in the long run, it is Edison's name that has survived, at the time his campaign damaged his own reputation more than Tesla's. He backed off.2
The detail that turned the campaign against its author is technical and it is horrible.
Edison's electrocution experiments had all been conducted on small creatures. The charge calibrated from those experiments was too weak for a man.
the man was only half killed. In perhaps the country's cruelest state-authorized execution, the procedure had to be repeated. It was an awful spectacle.1
That is the mechanism of the failure, and it is worth being exact about why it inverted.
Edison's argument was: AC is so dangerous it should not be in your home. The demonstration was designed to prove lethality.
What the demonstration actually proved was that AC, applied by Edison, failed to kill a man on the first attempt. The spectacle undercut the technical claim and simultaneously supplied the public with an image — a botched execution, repeated — that attached to the person who had arranged it rather than to the person whose system it was.
He had built the demonstration, chosen the venue, and lobbied for the occasion. There was no version in which the audience wondered whose idea it had been.
Greene's extracted lesson is the general form:
never go too far in attacks like these, for that will draw more attention to your own vengefulness than to the person you are slandering.2
The word vengefulness is precise, and it identifies the actual mechanism.
An attack on a rival's reputation asks the audience to evaluate a claim about the rival. But it simultaneously, and unavoidably, asks them to evaluate the attacker's conduct in making it — and those two evaluations compete for the same attention.
A proportionate attack keeps the audience looking at the claim. A disproportionate one moves them to the conduct, because the conduct has become the more remarkable thing in the room.
And once the audience is looking at the conduct, the original claim is not refuted — it is orphaned. Nobody concluded AC was safe. They concluded Edison was a man who electrocuted pets, and that conclusion did its work independently of anything about current.
So the failure is not that the attack was untrue or that it missed. It is that the attack generated a more interesting subject than its target, and audiences follow the more interesting subject every time.
Note also the direction of the asymmetry Greene sets up elsewhere in this law. He argues that attacking a reputation is "a potent weapon, particularly when you have less power than he does. He has much more to lose in such a battle, and your own thus-far-small reputation gives him a small target."3
Edison was the powerful party. He had the larger reputation and therefore the larger surface, and he attacked downward — which is the configuration Greene's own rule says will hurt the attacker most. The rule was available and Edison is the case that proves it.
Worth recording as a structural fact about the book rather than about the men.
Edison and Tesla appear in The 48 Laws of Power three times each — here at Law 5, at Law 6 in Edison's refusal of a shared Nobel, and at Law 7 over the fifty thousand dollars and the AC patents.
Edison wins all three. Tesla loses all three. They are the same three events, told in three different chapters to prove three different laws.
Greene never connects them, and the reader encountering them separately would not notice that the book has quietly assembled a complete account of one rivalry across two hundred pages — one in which the man who behaves worst succeeds consistently, which is a finding the individual chapters do not draw.
Greene's own hedge is the most interesting thing on the page and he does not develop it:
Although, in the long run, it is Edison's name that has survived…2
So the campaign damaged Edison's reputation at the time, he backed off, and a century later he is the inventor who harnessed electricity and Tesla is — in 1998, when Greene was writing — a specialist's name.
Which means Law 5's central claim needs a qualifier the chapter never supplies. Reputation is the cornerstone of power and once it slips, you are vulnerable, and will be attacked on all sides4 — both true of Edison in 1890, neither true of Edison in 1931.
Reputation is not a stock that depletes permanently. It has a half-life, and the contemporary damage Greene describes was survivable because the thing that eventually carried Edison's name was the accumulated output, not the conduct.
That is a genuinely different model from the one Law 5 argues, and the counter-evidence is inside Greene's own sentence.
A competitor is winning on something and you can see the weakness in what they're selling.
Before you say anything publicly, run the proportion check, because it's the only variable that matters and it is not about whether you're right.
Ask what a neutral observer will find more remarkable: your claim, or the fact that you made it.
If the claim is technical, specific, and stated once — the claim is more remarkable, and the audience stays on it. If it is repeated, escalating, or theatrical, you become the story, and the story is about your motives rather than their product.
Edison's first move was arguably legitimate: demonstrate a safety concern. His problem was the second and third, which converted a safety argument into a campaign, and a campaign has an author in a way that a demonstration doesn't.
Then the direction check, which is Greene's own rule and which he watched Edison break: attacks travel better upward than downward. If you are smaller, you have less surface and the exchange costs your rival more than it costs you. If you are the incumbent, every attack you make is fought on your own reputation's terrain and you are the one with something to lose.
The practical residue: if you are ahead, do not attack. Ship. The only argument available to a market leader that doesn't cost more than it returns is a better product, and Edison had the resources to build one and spent them on this instead.
Stated flatly, because it is part of what this source is.
The passage reports pets captured and electrocuted, and a man killed by a botched electrocution Edison lobbied to arrange, in the register of a tactical lesson. The extracted moral is about efficiency — never go too far in attacks like these — where "too far" is measured by damage to the attacker's reputation.
No other cost is priced. The vault records the technique, the outcome, and the framing, without smoothing any of the three.
Strongest support. A dated, public, historically attested episode with a documented reversal — Edison backed off — which makes the causal claim about reputational blowback unusually testable for this book.
Tension — Greene's own hedge undercuts Law 5. "In the long run, it is Edison's name that has survived." The chapter argues reputation is the cornerstone of power and that damage leaves you attacked on all sides. Edison's damage was real, contemporaneous, and did not determine the outcome. Reputation appears to have a half-life the law does not model.
Tension — the case is the wrong shape for its own law. Law 5's advice on attacking reputations is aimed at the weaker party, who has a small target and everything to gain. Edison is the strong party attacking downward. Greene uses the case to illustrate a rule whose stated conditions Edison did not meet.
Tension — "irresponsible in promoting it" was not entirely wrong. AC at transmission voltages is more dangerous than DC at domestic ones; the safety question was real, and the engineering answer (transformers, insulation, standards) arrived later. Greene presents the claim as pure slander. It was a real concern pursued by indefensible means, which is a different thing and a more interesting one.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no citation for the pet electrocutions, the lobbying of prison authorities, or the reputational effect. This is among the most-mythologised episodes in the history of technology and Greene gives the popular version without qualification. [POPULAR SOURCE].
Open questions.
Greene's Reputation as Controllable Fiction — the Law 5 mechanism page rebuilt in this same cluster — holds that reputation is constructed, defended, and lost catastrophically.
Edison is the case that shows the model is missing time.
The mechanism page treats reputation as a state: solid or slipped, guarded or breached. Greene's own hedge describes something else — a reputation that was damaged, survived the damage, and was eventually determined by an entirely different input a generation later.
What the pair yields: reputation as Greene models it is a short-run asset being used to explain long-run outcomes. The mechanism he describes is real and operates on a scale of months to years — Barnum destroying Peale's, Chuko Liang facing Sima Yi. Over decades a different process runs, in which accumulated output overwrites contemporaneous conduct, and Edison is the case where the two give opposite answers inside one paragraph.
→ behavioral-mechanics · Doubt as a Weapon
The technique page holds that an insinuation puts the target in a dilemma with no good exit — deny it and the defensiveness itself suggests substance, ignore it and the doubt stands unrefuted.
Edison is the case where the weapon discharges backwards, and the pairing shows exactly why.
Doubt works because it is cheap and deniable. A rumour has no author standing next to it; its power comes from circulating without a face attached, which is what makes the target's denial look disproportionate.
Edison did the opposite. He supplied the author, the venue, the animals and the occasion. Nothing was insinuated — everything was staged, publicly, by a named man with an obvious commercial interest.
The insight neither page produces alone: doubt and demonstration are opposite instruments, and demonstration is the weaker one against a reputation. A demonstration can be examined, and examining it means examining who built it. Doubt cannot be examined, which is its whole power. Edison had the strongest available weapon — a widely-held, plausible safety concern — and converted it into the weakest by attaching himself to it in public.
→ business · Attention as the Last Moat
The business page holds that attention is the scarce, defensible asset and that everything else follows from capturing it.
Edison captured an enormous amount of it and the capture destroyed the campaign, which is a case the moat thesis needs to account for.
The failure is directional. Attention is not a quantity you accumulate — it is a beam that lands somewhere, and the person who generates a spectacle does not control where. Edison generated the most attention available in 1890 and it landed on him rather than on the current.
The insight neither field states alone: attention captured by an attack always illuminates the attacker, because an attack is an act and acts have authors. Which is a structural asymmetry between building and attacking as attention strategies — a product that captures attention directs it at the product, while an attack that captures attention splits it between the claim and the conduct, and the conduct is the more human and therefore more compelling half.
Greene's "more attention to your own vengefulness than to the person you are slandering" is that split, described from the losing end.
Sharpest implication. The campaign failed not because it was untrue and not because it missed, but because it generated a more interesting subject than its target. Nobody watching a botched execution concluded that alternating current was safe; they concluded something about the man who had arranged it, and that conclusion ran independently of the technical claim forever after. An attack does not compete with its target — it competes with itself, and the audience picks whichever half is more remarkable.
Generative questions.