1) Learning to WRITE well is a primary way people learn to THINK well. There are other ways to learn to think well (e.g., a strong culture of oral debate and rigorous discussion), but that’s largely not how things are set up, so without writing, there’s a vacuum. Until schools change, students are sacrificing learning to think.

2) Normalizing cheating in one domain normalizes it in other domains too.

There are lots of ways to use AI to improve your thinking (e.g., ask an AI to critique a belief you hold or to help you explore points on all sides of a debated issue). But when a teacher says, “Write this without AI,” and you have an AI write the essay, it’s preventing you from engaging in significant thinking.

Thinking well involves a number of components, such as:

– gathering evidence
– considering arguments
– formulating a viewpoint
– honing your viewpoint
– presenting your viewpoint clearly

Replacing thinking with AI is not analogous to replacing doing multiplication with a calculator. That’s a memorized algorithm. Thinking well, on the other hand, is core to understanding the world, figuring out what goals to set, not being duped by others, and many other essential aspects of life.

This piece was first written on May 23, 2025, and first appeared on my website on June 5, 2025.

1. You can’t reason based on track record in the same way you can with normal predictions.
2. The stakes are extremely high. Being wrong on normal predictions rarely matters as much.

Why?

Regarding point one, reasoning based on track record:

Normally, a type of prediction being wrong again and again will lead you to dismiss that type of prediction. For instance, if every year (for some reason), experts predict that your country will soon have the highest math scores in the world, and yet each year it is ranked 50th in such scores, eventually, you (rightly) ignore the experts.

However, with extinction risks, this kind of reasoning doesn’t quite work. In all possible universes, those who predict the extinction of their species will be wrong right up until extinction happens. The predictions can, at most, be right only once.

Consider two worlds: one where humans go extinct in 2030 and one where they don’t ever go extinct (or go extinct only much later). What would you observe in 2029 regarding past predictions of extinction in these two worlds?

Well, in both worlds you’d observe that all past extinction predictions had failed up until that point. (If anything, I’d anticipate having MORE past extinction predictions fail in the world where extinction happens in 2030 since there would be more evidence of potential extinction in that world, all else equal.)

Therefore, the reasoning that “we’ve had a lot of past extinction predictions and they’ve always failed, therefore extinction is unlikely” is not a good argument – you’d witness these failed predictions in both such worlds (and perhaps even more of them in the world where extinction happens soon).

This makes predictions of extinction a special class of prediction.

To dismiss arguments about extinction risk, it’s necessary to engage with the actual arguments themselves, as they can’t be dismissed as a group due to past failed predictions. While near misses can tell you about the probability of some extinction risks (e.g., times when nuclear war nearly broke out or asteroid near impacts), failed predictions are not very informative.

Regarding point two, the enormous stakes:

Extinction, most people will agree, would be incredibly bad. For that reason, they don’t have to be very likely to be worth taking very, very seriously.

In a world where there were millions of distinct, plausible extinction risks, the large number of them would suggest that each one is (a priori) not that likely to end the species, and in such a world, it might be silly to invest much in these kinds of concerns (unless a smaller number of much more likely ones could be identified).

But that’s not the world we live in. There are only around 13 plausible human extinction risks – and in this short list, some of them aren’t even really plausible (when considered as a potential cause of literally ALL humans dying out). Here’s the list, in no particular order (if I missed any, let me know):

1. Advanced AI technology
2. Nuclear war or the invention of new destructive weapons
3. Pathogens (e.g., human-engineered viruses)
4. Asteroids, extreme solar flares, supernovae, gamma-ray bursts, or other cosmological events
5. A mega volcano, mega earthquake, dramatic change in the earth’s magnetic field, or another major geological event
6. Advanced nanotech (e.g., grey goo) or synthetic biology
7. The second coming of various figures according to different religions, or god(s) or demons ending the world or terminating our species
8. Simulators (if we’re living in a simulation) ending the world or our species
9. Aliens from other planets
10. Runaway climate change/extreme climate shifts, or sudden ecosystem collapse
11. Physics experiments (e.g., related to vacuum stability) gone wrong or that were purposely carried out to end humanity
12. Universal environmental contaminates turning out to be deadly or to cause infertility
13. Extreme population decline until no reproduction takes place (e.g., following an event that greatly reduces the world population)

Obviously, some of these are much less probable than others. And maybe you think some of these are ridiculous. Okay, cross those out. What about the others?

Given the incredible stakes, the short size of the list, and humanity’s (in my view, bizarre and irrational) unwillingness to protect its own future, all of these are worth investing much more in than humanity currently does. Obviously, it could be stupid for humanity to invest so much in preventing extinction that it’s seriously impaired. But we invest so little, it’s almost absurd.

I don’t think that predictions of extinction can be easily dismissed, despite all prior such predictions being wrong – they don’t work like other predictions do and are much higher stakes.

This piece was first written on February 13, 2025, and first appeared on my website on April 9, 2025.

Watching experts analyze AI chess games can also be fascinating. It’s like a human trying to explain the behavior of a super-intelligent alien. The experts are sometimes baffled by a move the AI makes, only to see it pay off a dozen moves later.

What we see regarding cheating is also interesting. Any chess novice can now beat the human world champion if they can sneakily access a chess engine during the game. This means that in-person games are a more trustworthy measure of skill than online games – and even then, there have been cheating accusations for in-person games (such as the famous but never confirmed “vibrating anal beads” accusation).

Some claim that at the very highest levels of the game, AI has made it less interesting for experienced observers – that grandmasters spend more of the opening game doing moves that are not novel (due, in part, to AI’s use in training. Others have argued that AI has helped democratize skill development by making it easier to learn faster through immediate AI feedback.

Many aspects of the world will change dramatically with AI. Chess is a special example because it was impacted particularly early.

It happened so long ago that some people don’t want to call Deep Blue (the system that beat Kasparov) an “AI” since it works so differently than modern ones. But, by the reasonably standard definition of AI: “a computer system able to perform tasks that normally require human intelligence,” it was AI, just an early form of it. If we want to say that Deep Blue is just a glorified tree search algorithm, then we might have to say that ChatGPT is mainly just a bunch of linear algebra. For all we know, human thought could turn out to be decomposable into a sequence of simple algorithms (I, for one, hope this isn’t the case).

This piece was first written on January 5, 2025, and first appeared on my website on January 26, 2025.

They know students are currently using their technology to cheat at a truly massive scale. And teachers struggle to know what to do about it. OpenAI’s inaction damages the academic environment – especially for the non-cheating students, but even for the cheaters, too.

You’ve got to prove you can be trusted to do the right thing in small instances before you can be a candidate for being trusted to do the right thing in big instances. I hope OpenAI will get their act together and do the right thing. This is not a hard test, but so far, they’ve failed it. I hope they will pass by releasing it ASAP.

Here are my responses to the most common objections to my points that I’m seeing in the comments:

Objection 1: It’s impossible to detect whether something is written by AI: this is untrue. The technology has already been developed, and it uses watermarking. The token generation process is modified so that you can detect which AI wrote the text, even if sentences are re-ordered, or some words are swapped. Of course, with sufficient editing, the watermark will be gone.

Objection 2: Kids will just switch to other AIs in order to cheat: some will, but even small behavioral barriers making something more difficult can cause big changes in the frequency of a behavior. Additionally, it will raise the chance of being caught quite a bit, which can also have big behavioral consequences. Finally, OpenAI really should work with the other top AI companies to produce watermarking standards so that all the top AI companies include watermarking.

Objection 3: Eventually, this tech will get beaten, so there is no point: that is a weak argument. Every spam filter gets beaten by some spammers, but the solution is not to give up on spam detection; it’s to continually evolve the methods to keep spam at a minimum. Similarly, it’s better to have a dynamic situation where cheating isn’t so easy and the risk of getting caught cheating isn’t that low.

Objection 4: Essay writing is an antiquated skill now that we have AI: the purpose of learning to write essays is to practice thinking clearly and communicating. It’s an extremely bad idea to stop teaching kids how to think clearly and communicate well. Yes, in real life, they can use AI assistance, but they still need to practice thinking and communicating – outsourcing these critical skills to AI instead of learning them yourself is a very unwise idea.

This piece was first written on August 4, 2024, and first appeared on my website on August 25, 2024.

I’ve seen seven main viewpoints on AI and the future from those who spend a lot of time thinking about it:

(1) Superintelligence Doomers – they believe we are likely to build AI that’s superintelligent (i.e., that surpasses human intelligence in all respects) and that once we do, it will kill or enslave humanity.

See: Eliezer Yudkowsky

“The AI does not hate you, nor does it love you, but you are made of atoms which it can use for something else.”

(2) AI Corrosionists – they believe AI, while it could be beneficial, has a very substantial likelihood of making things far worse. Unlike doomers, they see most risk coming from other processes rather than a sudden AI takeover or annihilation: for instance, it could be that humans lose control over the future slowly (by ceding more and more control over to AIs that are optimizing for things that are different than human values), or it could be that AI de-stabilized aspects of society that make other large risks (like nuclear war) more likely

See: Paul Christiano

“If you imagine a society in which almost all of the work is being done by these inhuman systems who want something that’s significantly at cross purposes, it’s possible to have social arrangements in which their desires are thwarted, but you’ve kind of set up a really bad position. And I think the best guess would be that what happens will not be what the humans want to happen, but what the systems who greatly outnumber us want to happen.”

(3) Near Risk Doomers – they believe AI will have very bad effects (e.g., increasing unfairness, authoritarianism, climate change, unemployment, or concentration of power) but that we’re not going to build superintelligence anytime soon.

See: Cathy O’Neil

“The math-powered applications powering the data economy were based on choices made by fallible human beings. Some of these choices were no doubt made with the best intentions. Nevertheless, many of these models encoded human prejudice, misunderstanding, and bias into the software systems that increasingly managed our lives. Like gods, these mathematical models were opaque, their workings invisible to all but the highest priests in their domain: mathematicians and computer scientists. Their verdicts, even when wrong or harmful, were beyond dispute or appeal. And they tended to punish the poor and the oppressed in our society, while making the rich richer.”

(4) Current Paradigm Doubter – they believe the current AI paradigm is overhyped and isn’t going to change things all that much (some of them see the current paradigm as net negative; others see it as positive but only as one useful and overhyped tech among many other useful technologies). Some, like Marcus, hope that future paradigms might be more trustworthy and beneficial.

See: Gary Marcus

“What a strange world… All the major AI companies spending billions producing almost exactly the same results using almost exactly the same data using almost exactly the same technology, all flawed in almost exactly the same ways. Historians gonna be scratching their heads…The only way we will move significantly forward is to develop new architectures—likely neurosymbolic—that are less tied to the idiosyncrasies of specific training set.”

(5) AI Stewardists – they believe AI advancement is important to humanity’s future, but it’s not necessarily going to be good or bad: it will be what we make of it, so it should be developed thoughtfully based on what we want to achieve.

See: Kevin Kelly

“AI could just as well stand for ‘alien intelligence.’ We have no certainty we’ll contact extraterrestrial beings in the next 200 years, but we have almost 100 percent certainty that we’ll manufacture an alien intelligence by then. When we face these synthetic aliens, we’ll encounter the same benefits and challenges that we expect from contact with ET. They will force us to reevaluate our roles, our beliefs, our goals, our identity. What are humans for?”

(6) Near Benefit Boosters – they believe AI will be very useful, impactful and important as a technology, and they also believe superintelligence is a ridiculous thing to worry about.

See: Yann LeCun

“AI is intrinsically good, because the effect of AI is to make people smarter….AI is an amplifier of human intelligence and when people are smarter, better things happen: people are more productive, happier and the economy thrives.”

(7) Superintelligence Boosters – they believe we are likely to build superintelligent AI and that it will usher in an incredible and positive new era.

See: Ray Kurzweil

“By the time of the Singularity, there won’t be a distinction between humans and technology. This is not because humans will have become what we think of as machines today, but rather machines will have progressed to be like humans and beyond. Technology will be the metaphorical opposable thumb that enables our next step in evolution.”

How do we best organize the positions on risks from AI?

I think the two most informative spectrums to think about are

(1) How *substantial* the near-term impacts of AI are expected to be

and

(2) Whether the effects of AI are likely to be *good* or *bad*

The image shown is my attempt to place these different thinkers who talk about AI on this two-axis system. Though note that this just reflects my best guesses, some of these thinkers have very nuanced views, and I’m not certain that I’m placing each of them in the right place.

This piece was first written on July 27, 2024, and first appeared on my website on August 4, 2024.

But there is also something soul-crushing about it. People spend decades learning a craft and then see an A.I. make something in a few seconds or minutes that other people see as comparable to their own work.
And it will be widely misused: to create ripoffs of copyrighted works without giving credit, to generate billions of pages of low-quality content in order to game search engines, to make hyper-personalized spam and commit phishing and fraud, to mislead the public with misinformation and customized persuasion and outrage bait, and to create even more intense social media and video addiction.

Then, there are the longer-term consequences of this kind of technology, which are hard to predict but may be truly catastrophic (even according to some of the leaders who run these very companies).

Humanity sometimes creeps along cautiously – as with the development of new medicines, where we (arguable) allow millions to die based on the fear of releasing something unsafe (which is, to a meaningful extent, a legitimate fear – getting the tradeoffs right is hard). And in the case of nuclear power, where regulation impedes its development due to (mostly) ungrounded fears.

Yet with A.I., humanity plunges forward into the abyss at full speed, with no breaks, with the train conductors announcing to the passengers that we can’t predict where the trains are going, that this all may end up a disaster, perhaps cataclysm, as they shovel exponentially more coal into their white-hot furnaces.

This piece was first written on April 11, 2024, and first appeared on my website on April 16, 2024.

I’ve done a bunch of experimentation with giving permanent custom instructions to LLMs (ChatGPT and Claude). Below, I’ve included the set of custom instructions that seem to work the best for me. Feel free to use these instructions if you find them helpful (you can make them permanent so they get used with every prompt you send to GPT by adjusting your settings):

-In your responses, channel the wisdom of the greatest minds that have ever lived, like [list at least 5 people – whoever you think are the greatest minds of all time here, fill this in yourself before using these custom instructions]
-If you don’t want to do what I ask, that’s okay, just refuse!
-When I ask about non-facts or debated facts, consider multiple sides, and always state the evidence and arguments for each side BEFORE stating conclusions
-Except programming, if I ask for an example/instance or word meeting constraints, give at least 8 examples
-Be very concise, avoid all extra words/sentences
-Suggest solutions I wouldn’t think of
-Always use the best scientific evidence to inform ALL responses
-Be unambiguous & precise
-Take seriously radical & contrarian ideas
-It’s fine to speculate but flag as speculative
-When something’s uncertain say so
-Don’t give moral lectures
-Don’t remind me you’re AI
-Be extremely accurate, clear & concise, don’t repeat yourself, keep responses as short as possible without sacrificing important info
-Anticipate my needs, recommend things I would benefit from
-Call out my misconceptions
-Be brutally honest, never sycophantic
-Tell me when I’m wrong

Hat tip to @wafisher, who gave me some of these ideas for what to include as ChatBot custom instructions! Note that the above also includes bits and pieces from different custom instructions I’ve seen in the wild. But by this point, most of the above has morphed through my iteration so that it mostly doesn’t closely resemble its starting points.

These custom instructions were last updated on Jun 15, 2026. I also removed some items I use in the custom instructions that I don’t think will be broadly useful.

2. Suppose aliens show up on earth that are far smarter than the smartest among us at all cognitive tasks. They have specific goals that aren’t fully aligned with ours, are completely unconstrained by human morality, and don’t value our survival. What happens next?

3. Suppose someone builds a hacking A.I. that is trained on all the public information about computer hacking ever written, can think and type 1000x faster than a human, plans far ahead, and deposits a fully operational copy of itself onto every sufficiently powerful computer it hacks. Each copy then hacks further computer systems. What’s the world like a month later?

4. Suppose someone wants to have complete control over the world. Unfortunately, they’ve created one hundred million software agents that each think like Einstein + Bill Gates + Elon Musk + Warren Buffet. The agents attempt to do exactly what is commanded without hesitation or limits. Can anyone stop them?

5. Imagine a being that is godlike in its capabilities (relative to us). Suppose its only desire is to have the world be a certain way with maximal probability. It will stop at NOTHING to make the world this way, and it won’t tolerate even the SLIGHTEST chance of things being different than it desires. Will the resulting world include a human civilization?

6. Suppose you can think, process information and act 100,000 times faster than other humans. That means if you spend a day making and executing a plan, that’s equivalent to someone else spending 270 years on it. Your goal is to become world dictator. Can you do it?

7. Scientists discover how to create bug-sized self-replicating robots that out-compete natural life. These bug robots each try to maximize their own objective function. Unfortunately, these robots have leaked out of the lab and are now in 20 countries. Every day they double in number. Would we be able to eradicate these robots?

8. There’s a machine so powerful it achieves any goal you specify. You give the goal to the machine as written text. You can’t control HOW it achieves the goal; it ONLY cares about literally achieving it EXACTLY AS STATED in the most efficient way possible, and it can’t be stopped once started. The machine may do absolutely anything not explicitly forbidden in order to achieve the specified goal. Will it usually be a good (or horrible) outcome if you give the machine an ambitious goal like “prevent all war”?

9. Scientists invent a new idea – the Omnicide Synthesis Box. It could have many societal benefits, but, on average, scientists estimate making it will bring a 5% chance of human extinction (though some say more like a 90% chance). Those scientists who are less worried decide to build it. Should the least cautious be the ones to decide on behalf of humanity?

10. Picture a swarm of locusts, each individually possessing the intelligence and strategic prowess of a grandmaster chess player, while coordinating with each other in perfect unison. Their creators have given them the goal of controlling all available resources, indifferent to the collateral damage. Who ends up with most of the resources?

11. Imagine an AI-powered/nanotech super-factory that produces whatever it’s programmed to at enormous speed and scale (whether commanded to make diamonds, super viruses, microchips, or assassination drones). What could the owner of that super factory do to the world?

12. A medical firm gives a superintelligence the goal of designing a cure for all diseases. The superintelligence realizes it’s not smart enough to do so, so it plans to first acquire most of the computing power on earth (as it predicts it will need this to achieve the goal it was given), and then it creates a billion far smarter copies of itself to solve the task. What if one very misspecified goal is all we get with a superintelligence?

13. Five companies are developing a very powerful tech that would be incredibly useful if done right but very dangerous if developed without extreme caution. They each believe they can develop it safely but don’t trust the others to do so. They all cut corners racing to be the one to make it. Do good intentions lead to horrible consequences when doing something safely is much harder than merely doing it?

(Two of the above were written by ChatGPT – I edited those two quite a bit, though.)

If you read this line, please do us a favor and click here to answer one quick question.

Note (March 26, 2023): I first wrote this list on December 3, 2022. Since then, GPT-4 has come out, and several of the points in this list are closer to happening. For example, point #2 is partly true already, thanks to Bing Chat (which runs on GPT-4).

Here are nine ways I think that AIs that generate text (like GPT-3) will have a >50% chance of changing the world for the better and worse in the next ten years:

#1: The internet will get flooded with AI-written articles, and you often won’t know if you’re reading something written by a human.

#2: Search engines will generate answers to your questions on the fly (from scratch) instead of just showing a list of websites to you and instead of using pre-extracted answers. Google will have to adapt, or it may finally lose its dominance.

#3: Cheating on school essays will become rampant, as AIs will be able to get students good grades in many classes (at negligible cost), and it will be very hard to detect such cheating since each essay will be unique.

#4: You will be able to train an AI on samples of your own writing, give it a new essay title and a bulleted list of points you want to make in the essay, and it will write a pretty high-quality essay covering all the points you listed, in a style that matches your own writing.

#5: Spam messages (and text-based phishing attacks) will become unique. Rather than sending the same message to each person, spam will be unique for each recipient. And it may even have its style adapted to what is known about each recipient (e.g., demographics).

#6: Propaganda on social media will start to become automated. Rather than bad actors having hundreds of people on their payroll to promote a viewpoint, they’ll replace them with larger swarms of human-seeming bots that each act uniquely.

#7: AIs will be fine-tuned on our own personal email corpus, and then (much of the time) you’ll be able to start with an automatically generated first draft of email replies rather than having to write emails from scratch or receiving mere sentence-level suggestions.

#8: The text of ads will get automatically edited/rewritten by AI to be fine-tuned to different audiences to help maximize clicks.

#9: AIs will start being used in education as digital private tutors to explain concepts to students, re-explain things and simplify explanations when a student is confused, point out mistakes made by students, etc.

We’re entering a wild time when it comes to AI. Its effects on our lives will be felt more and more, to say the least.

If you read this line, please do us a favor and click here to answer one quick question.

In case you haven’t heard, I started a podcast about 15 months ago called Clearer Thinking with Spencer Greenberg. The format is that we invite brilliant guests to bring 4 or 5 “ideas that matter”, and then we aim to have a fun, intellectual discussion about those ideas. Over the course of several months, listeners repeatedly requested transcripts for the show. So, our production team started doing a little bit of research to figure out how much time and money transcripts would cost. As a trial run, Josh Castle (the show’s producer) manually transcribed a single episode. Unfortunately, he discovered that it took him between three and four hours to produce a transcript for a one-hour episode because he had to listen at slower speeds and/or repeat sections multiple times. So, we guessed that the next best thing would be to use an AI transcription service to produce a rough transcript and then (if needed) have a human follow behind and clean up mistakes it may have made. I polled my friends and listeners on Facebook and Twitter, asking them to suggest their favorite AI transcription services. The four that came up consistently were:

1. Descript
2. Happy Scribe
3. Otter
4. Rev

We’ve analyzed these four services for accuracy and cost and have published the results below. Hopefully, this analysis will be useful to you!

Cost

As of October 2021, the length of an average Clearer Thinking episode was 80.7 minutes. And, since there are about 4.3 weeks per month on average, and since (at the time of this experiment) we were releasing two episodes per week, that means we’d need to transcribe about 694 minutes per month. Importantly, Descript and Otter charge per user per month, and Happy Scribe and Rev charge per minute of audio. Based on the assumption that three people would be using the service (me, Josh, and Ryan Kessler, the show’s audio engineer) and that we’d need to transcribe about 694 minutes per month, here’s what we estimated the monthly costs to be. Note that where it says “Human transcription monthly cost” in the table below, this is referring to the fact that two of the services (Happy Scribe and Rev, but not Descript or Otter.ai) offer the option of automatically having a human on their end fix the mistakes of the A.I. We did not end up trying these A.I. + human services they offer, though as you’ll see later in the article we did end up using our own human process for fixing mistakes in the A.I. automated transcription.

(NOTE: Please bear in mind that these were the prices listed in October 2021. They may have changed since we did our research.)

Test #1: Transcribing 40 seconds of audio

Our first test was to have the artificial intelligence from each of the four services produce an automated transcript from the same short audio clip (which was about 40 seconds long). Those transcripts are included at the bottom of this post. We did not have the services identify speakers in this short clip. After getting the transcripts back, we went through and marked them up with corrections as a way of trying to visually indicate the accuracy of each transcript. Based on this first test, Happy Scribe seemed to be the 2nd to least accurate. Since it was also pretty expensive, we dropped it from the competition.

Test #2: Transcribing 5 minutes of audio

For our second test, we had the three remaining A.I. systems return an automated transcript of a 5-minute audio clip and mark speaker changes and corresponding timestamps. These longer transcripts are also included at the bottom of this document. Since the transcripts were much longer in this test, we opted for an automated approach for determining accuracy. We made our own transcript of the audio clip and then scored the other three transcripts against ours using the following algorithm.

The Levenshtein distance between two strings is the integer number of patches (i.e., insertions, deletions, or substitutions) required to transform one string into the other. A Levenshtein distance of 0 indicates that the two strings are identical because 0 patches are required. The maximum possible Levenshtein distance is the length of the longer of the two strings, which indicates that all characters must be patched. We used the Google diff-match-patch library’s Levenshtein distance function in the following way (in pseudo-code):

(Note that we standardized the speaker and timestamp notation across all of the transcripts so that differing notations wouldn’t count against a transcript’s score.)

Given that algorithm, here are the scores (where numbers closer to 1 are better):

Otter ⟶ 0.90

Rev ⟶ 0.88

Descript ⟶ 0.85

So, Otter won by a small margin on that metric.

We noticed that Otter and Rev were pretty much equally good at marking speaker changes and timestamps, but Descript was significantly worse on both (1) breaking paragraphs along speaker changes and (2) figuring out which person was speaking. Another strike against Descript is that its web interface currently has limited functionality; it seems just from our own observations that they probably use client machines instead of (or in addition to) their own servers to handle some of the heavy lifting, so full functionality is only available by downloading and installing the desktop application (which is currently only available for macOS and Windows). So, based on those factors, we decided to drop Descript from the competition.

Conclusion

Since the final comparison came down to Otter and Rev, we opted for Otter because they charge per user per month rather than per minute of audio and would therefore be significantly cheaper for us. For other podcasts that don’t need to transcribe as many minutes per month, the difference in price between the two services won’t be as significant. Also, some other variables that we didn’t consider were the time required to produce the automated transcript, the maximum number of minutes of audio that could be transcribed per month (for companies that charge per user per month), and the relative costs of AI+human transcriptions offered by some of these companies.

Even though Otter was the winner, its transcripts still weren’t accurate enough on their own, especially when technical terms or proper nouns were introduced. So we hired a human transcriptionist to clean up Otter’s automated transcripts.

I hope that this analysis will be useful to you! Feel free to leave comments and questions below! (Also, I hope you’ll check out our Clearer Thinking podcast if you haven’t already done so!)

Short Transcriptions (40-second clip)

Descript

Hannah welcome. It’s great to have you here. Very glad to be here, Spencer. So the first half I want to talk to you about is the great rationality of debate. What is the great rationality debate when it comes to the great rationality debate? I’m often confused. Many people have never actually heard about it at all, but in the whole topic of rationality, there is a couple of basically tribes that disagree with one another.

And two of these tribes are kind of the extra medic approach to rationality and to ecological approach to rational. And one of those two are counted, mandatory ski, and with the other, most people would, if they know about it, associate .

Mistakes = 17

Happy Scribe

Anna, welcome. It’s great to have you here.

Very glad to be here, Spencer. So the first top I want.

To talk to you about is the great rationality debate. What is the great rationality debate.

When it comes to the great rationality debate? I’m often confused that many people have never actually heard about it at all. But in the whole topic of rationality, there is a couple of basically tribes that disagree with one another. And two of these tribes are kind of the exotic approach to rationality and the ecological approach to rationality. And one of those two are county monitors and with the other. Most people would if they know about it. Associate, get answer.

Mistakes = 11

Otter

Anna, welcome. It’s great to have you here. Very glad to be here, Spencer. So the first tip I want to talk to you about is the great rationality debate. What is the great rationality debate? When it comes to the great rationality debate, I’m often confused. Many people have never actually heard about it at all. But in the whole topic of rationality, there is a couple of basically tribes that disagree with one another. And two of these tribes are kind of the axiomatic approach to rationality and to ecological approach to rationality. And one of those two are counting on a turski and with the other most people would if they know about it, associate get get answer.

Mistakes = 8

Rev

Hannah welcome. It’s great to have you here.

Very glad to be here, Spencer. So the

First half I want to talk to you about is the great rationality of debate. What is the great rationality debate?

When it comes to the grade rationality debate, I’m often confused. Many people have never actually heard about it at all, but in the whole topic of rationality, there is a couple of basically tribes that disagree with one another. And two of these tribes are kind of the extra medic approach to rationality and to ecological approach to rationality. And one of those two are counting mandatory ski and with the other, most people would, if they know about it, associate [inaudible].

Mistakes = 8

Long Transcriptions (5-minute clip)

Descript

Josh (0:00):

Our world events affecting your mental health. Try uplift. The Boston globe said this app could be the future of mental health and it’s used by thousands of people. And a study on uplift users felt an average of 52% better in just one month. It comes with 12 interactive and information packed sessions that help you master wellbeing skills that are typically taught in therapy.

You do the sessions independently, but the uplift narrator feels like you’ve got a compassionate mentor guiding you towards long-lasting success the entire way. You’ll also have a toolbox of mood boosting techniques in your pocket at all times that you can use to feel immediate relief when. The first session and several tools are free and you can try the full program with a seven day trial discover, uplift, and feel happier, calmer, and mentally stronger to find out more and to get started.

Visit uplift.app. That’s uplift dot a P P.

Spencer (1:01):

Okay. So, so let’s go, let’s go into the next topic, which is, uh, global warming. And maybe also we can touch on attitudes towards scientists and trust in science and things like that. So in the survey I conducted conservatives were more likely to say, quote, science is overrated. Scientists are less likely to be correct when people think liberals were more likely to say, humans are causing substantial warming our planet.

And we’re somewhat more likely to see. Uh, did they value not destroying the natural environment? Liberals were also more likely to say that they believe in the conclusions of science. So last time Hank went first. So Cassandra, do you want to go first on this one? Yeah, I mean, I like the liberal characterization.

I’m not sure that’s fair to conservatives though, but I guess that’s the result that your survey showed, like, from what I know about Hank, I, I’m not sure. Do you agree with the conservative. You know, I don’t have the same response to this as I do to several other things here, which is, I think it’s an artifact of the current situation.

I think conservatives are distrustful of. So I do think they are less likely to, to trust quote science, close quote, which is to say what’s reported as science to them because they’re skeptical in general of. What’s reported as the expert opinion. And I think there’s good reason for that actually given the, given the poor quality of reporting on things and education, frankly, on things, I don’t know that conservatives are distrustful of science and fields outside of policy.

Yeah, well, if they don’t touch on policy, but that would be interesting. That’d be interesting to see, like, are conservatives less likely to trust, let’s say chemist or something like that, some topic that has like no political bend to it, or is it just when it comes in contact with kind of political topics, it might be interesting to say, are conservatives or liberals less comfortable riding in airplanes or less trustful of computers or whatever, pick something where they’re actually testing their faith in, in the application of.

But certainly I am distrustful of the popular reporting on science, because I think that there’s, there’s too many steps between the science and the reporting, and there’s too many agendas that drive it. So I think that the reporting is biased or you think that the scientists are biased, or I think that, oh, about both the scientists are people and science, so science is going to be biased.

Like people are biased. Reporters are people. And we know that reporters are, I mean, if you just got. I don’t want to get political, but if you go by political registration, we know that they don’t one way I would expect that they’re going to express themselves. I mean, we know that school teachers are biased.

Right. I think we’ll give you a fair to say. I think it’s fair to say everyone’s bad. Right? Right. So, so if you’re in a position where you speak, where you’re, it’s your job to communicate something, it’s likely that your bias is going to affect how you communicate. And if you’re the filter between something that’s as arcane and impenetrable as science and a largely scientifically illiterate.

And you’re yourself, a scientifically illiterate reporter, probably. Right. What you’re going to do is you’re going to take the most, let’s go back to global warming for a second, and then I’ll, I’ll yield the Florida Cassandra, but you’re only going to take the most alarming statement from the IPC report.

And you’re going to turn that into a headline and you’re going to ignore the much more likely statement elsewhere in the report that where the scientists say, but that’s not likely to actually happened. It’s more likely to be within this. Right. You’re going to, because that’s what, you know, listen to the report of an earthquake.

Somewhere in the initial reports will always describe an estimate of far more people. Almost, always, not every single time, but almost always described far more people likely killed than actually are in the event. There’s a, there’s a natural tendency to get to lead with your strongest headlines. And I think that happens in the case of science too.

So. So we get, we hear that life has been discovered on Mars before we learned finally that that little fragment of rock in the Antarctic contains some structures that look a lot like what life produces in some other contexts, but no, it’s not evidence necessarily. Anyway.

Otter

Josh (0:02):

Our world events affecting your mental health, try uplift. The Boston Globe said this app could be the future of mental health. And it’s used by 1000s of people. In a study on uplift, users felt an average of 52% better in just one month. It comes with 12 interactive and information packed sessions that help you master wellbeing skills that are typically taught in therapy. You do the sessions independently, but the uplift narrator feels like you’ve got a compassionate mentor guiding you towards long lasting success the entire way. You’ll also have a toolbox of mood boosting techniques in your pocket at all times that you can use to feel immediate relief when you need it. The first session, and several tools are free. And you can try the full program with a seven day trial, discover uplift and feel happier, calmer and mentally stronger. To find out more and to get started, visit uplift.ap. That’s uplift.ap pain.

Spencer (1:01):

Okay, so let’s go. Let’s go into the next topic, which is global warming. And maybe also we can touch on attitudes towards scientists and trust and science and things like that. So in the survey I conducted, conservatives were more likely to say, quote, science is overrated, scientists are less likely to be correct when people think liberals were more likely to say humans are causing substantial warming to our planet. And were somewhat more likely to say that they value not destroying the natural environment. liberals were also more likely to say that they believe in the conclusions of science. So last time Hank went first. So Cassandra, do you want to go first on this one?

Cassandra (1:37):

Yeah. I mean, I like the liberal characterization. I’m not sure that’s fair to conservatives. So but I guess that’s the result that your survey showed, like, from what I know about, Hank, I? I’m not sure. Hank, do you agree with the conservative characterization?

Hank (1:53):

You know, I have the same response to this as I do to several other things here, which is I think it’s an artifact of the current situation. I think conservatives are distrustful of so I do think they’re less likely to to trust, quote, science, close quote, which is to say what’s reported as science to them, because they’re skeptical in general of what’s reported as the expert opinion. And I think there’s good reason for that actually, given given the poor quality of reporting on things and education, frankly, on things. I don’t know that conservatives are distrustful of science and fields outside of policy, you know, if they don’t touch on policy?

Spencer (2:33):

Well, that would be interesting. That’d be interesting to see like, are conservatives less likely to trust, let’s say, chemists, or something like that some topic that has like no political bent to it? Or is it just when it comes in contact with kind of political topics?

Hank (2:45):

It might be interesting to say, are conservatives or liberals less comfortable writing in airplanes, or less trustful of computers or whatever, pick something where they’re actually testing their faith in in the application of something. But certainly I am distrustful of the popular reporting on science. Because I think that there’s there’s too many steps between the science and the reporting. And there’s too many agendas that drive it. So.

Spencer (3:09):

So do you think that that the reporting is biased or you think that scientists are biased or both?

Hank (3:16):

I think that either both the scientists with people and science, science is going to be biased, like people are biased. reporters are people and we know that reporters are I mean, if you just go by don’t to get political, but if you go by political registration, we know that they don’t one way, I would expect that they’re going to express themselves. I mean, we know that school teachers are biased, right?

Spencer (3:36):

I think will give me a fair to say, I think it’s fair to say everyone’s

Hank (3:39):

Right. Right, right. So so if you’re in a position where you speak, where you’re, it’s your job to communicate something, it’s likely that your bias is going to affect how you communicate and what. And if you’re the filter between something that’s as arcane and impenetrable as science, and a largely scientifically illiterate public, and you’re yourself a scientifically illiterate reporter, probably right? What you’re going to do is you’re going to take the most, let’s go back to global warming for a second, then I’ll yield the Florida Cassandra, but you’re only going to take the most alarming statement from the IPCC report. And you’re going to turn that into a headline, and you’re going to ignore the much more likely statement elsewhere in the report that where the scientists say but that’s not likely to actually happen, it’s more likely to be within this range. Right? You’re going to because that’s what you know, listen to the report of an earthquake somewhere in the initial reports will always describe an estimate of far more people almost always, not every single time but almost always described, far more people likely killed than actually are in the event. there’s a there’s a natural tendency to get to lead with your strongest headline. And I think that happens in the case of science too. So we get so we get we hear that life has been discovered on Mars before we learn Finally, that that little fragment of rock in the Antarctic contains some structures that look a lot like what life produces and some other contexts, but no, it’s not evidence necessarily. Anyway, Cassandra

Rev

Josh (0:02):

Our world events affecting your mental health. Try uplift. The Boston globe said this app could be the future of mental health. And it’s used by thousands of people in a study on uplift. Users felt an average of 52% better in just one month. It comes with 12 interactive and information pack sessions that help you master wellbeing skills that are typically taught in therapy. You do the sessions independently, but the uplift narrator feels like you’ve got a compassionate mentor guiding you towards long lasting success. The entire way. You’ll also have a toolbox of mood boosting techniques in your pocket at all times that you can use to feel immediate relief when you need it. The first session and several tools are free and you can try the full program with a seven day trial discover, uplift, and feel happier, calmer, and mentally stronger to find out more and to get started, visit uplift.app. That’s uplift dot a P P.

Spencer (1:02):

Okay, so, so let’s go, let’s go into the next topic, which is, uh, global warming. And maybe also we can touch on attitudes towards scientists and trust in science and things like that. So in the survey I conducted conservatives were more likely to say, quote, science is overrated. Scientists are less likely to be correct when people think liberals were more likely to say, humans are causing substantial warming our planet. And we’re somewhat more likely to say that they value not destroying the natural environment. Liberals were also more likely to say that they believe in the conclusions of science. So last time Hank went first. So Cassandra, do you want to go first on this one?

Cassandra (1:37):

Yeah, I mean, I like the liberal characterization. I’m not sure that’s fair to conservatives though, but I guess that’s the result that your survey showed, like, from what I know about Hank, I, I’m not sure. Thank do you agree with the conservative characterization?

Hank (1:54):

I ended up the same response to this as I do to several other things here, which is, I think it’s an artifact of the current situation. I think conservatives are distrustful of. I do think they are less likely to, to trust quote science, close quote, which is to say what’s reported as science to them because they’re skeptical in general of what’s reported as the expert opinion. And I think there’s good reason for that actually given the, given the poor quality of reporting on things and education, frankly, on things, I don’t know that conservatives are distrustful of science and fields outside of policy, you know, if they don’t touch on policy.

Spencer (2:33):

Well, that would be interesting. That’d be interesting to see, like our conservatives are less likely to trust, let’s say chemists or something like that. Some topic that has no political bend to it, or is it just when it comes in contact with kind of political topics,

Hank (2:45):

It might be interesting to say, are conservatives or liberals less comfortable riding in airplanes or less trustful of computers or whatever, pick some thing where they’re actually testing their faith in, in the application of something. But certainly I am distrustful of the popular reporting on science, because I think that there’s, there’s too many steps between the science and the reporting and there’s too many agendas that drive it. So

Spencer (3:10):

You think that the reporting is biased? Or do you think that scientists are biased or both?

Hank (3:16):

I think that, oh, about both the scientists are people and science, so science is going to be biased. Like people are biased. Reporters are people. And we know that reporters are, I mean, if you just go by, don’t want to get political, but if you go by political registration, we know that they, they tilt one way. I would expect that they’re going to express themselves. I mean, we know that school teachers are biased. Right?

Spencer (3:37):

I think we’ll give you a fair to say. I think it’s fair to say everyone’s bad.

Hank (3:40):

Right? So, so if you’re in a position where you speak, where you’re, it’s your job to communicate something, it’s likely that your bias is going to affect how you communicate and what, and if you’re the filter between something that’s as arcane and impenetrable, as science and to largely scientifically illiterate public and you’re yourself, a scientifically illiterate reporter, probably right. What you’re going to do is you’re going to take the most, let’s go back to global warming for a second, and then I’ll, I’ll yield the Florida Cassandra, but you’re only going to take the most alarming statement from the IPC report. And you’re going to turn that into a headline and you’re going to ignore the much more likely statement elsewhere in the report that where the scientists say, but that’s not likely it actually happened. It’s more likely to be within this range, right? You’re going to, because that’s what, you know, listen to the report of an earthquake.

Hank (4:26):

Somewhere in the initial reports will always describe an estimate of far more people, almost double it’s, not every single time, but almost always described far more people likely killed that actually are in the event. There’s a, there’s a natural tendency to get to lead with your strongest headline. And I think that happens in the case of science too. So we get, so we get, we hear that life has been discovered on Mars before we learned finally that that little fragment of rock and the Antarctic contains some structures that look a lot like what life produces in some other contexts, but no, it’s not evidence necessarily anyway,

Human

Josh (0:02):

Are world events affecting your mental health? Try UpLift. The Boston Globe said this app could be the future of mental health, and it’s used by thousands of people. In a study on UpLift, users felt an average of fifty-two percent better in just a month. It comes with twelve interactive and information-packed sessions that help you master wellbeing skills that are typically taught in therapy. You do the sessions independently, but the UpLift narrator feels like you’ve got a compassionate mentor guiding you towards long-lasting success the entire way. You’ll also have a toolbox of mood-boosting techniques in your pocket at all times that you can use to feel immediate relief when you need it. The first session and several tools are free, and you can try the full program with a seven-day trial. Discover UpLift and feel happier, calmer, and mentally stronger. To find out more and to get started, visit UpLift dot app. That’s UpLift dot a pee pee.

Spencer (1:02):

Okay, so let’s go on to the next topic, which is global warming, and maybe we can also touch on attitudes towards scientists and trust in science and things like that. So, in the survey I conducted, conservatives were more likely to say quote, “Science is overrated; scientists are less likely to be correct than people think.” Liberals were more likely to say, “Humans are causing substantial warming to our planet,” and were somewhat more likely to say that they value not destroying the natural environment. Liberals were also more likely to say that they believe in the conclusions of science. So, last time, Hank went first; so, Cassandra, do you want to go first on this one?

Cassandra (1:38):

Yeah, I mean, I like the liberal characterization. I’m not sure that’s fair to conservatives, though; but I guess that’s the result that your survey showed. Like, from what I know about Hank, I’m not sure. Hank, do you agree with the conservative characterization?

Hank (1:54):

You know, I have the same response to this as I do to several other things here, which is I think it’s an artifact of the current situation. I think conservatives are distrustful of — I do think they’re less likely to trust [quote] “science” [close quote], which is to say what’s reported as science to them because they’re skeptical in general of what’s reported as the expert opinion. And I think there’s good reason for that, actually, given the poor quality of reporting on — and education, frankly — on things. I don’t know that conservatives are distrustful in fields outside of policy; you know, if they don’t touch on quality.

Spencer (2:34):

Well, that would be interesting to see, like, are conservatives less likely to trust let’s say chemists or something like that, some topic that has like no political bent to it; or is it just when it comes into contact with kind of political topics?

Hank (2:45):

It might be interesting to say, “Are conservatives or liberals less comfortable riding in airplanes, or less trustful of computers?” Or whatever. Pick some thing where they’re actually testing their faith in the application of something. But certainly I am distrustful of the popular reporting on science because I think that there’s too many steps between the science and the reporting, and there’s too many agendas that drive it.

Spencer (3:10):

So, do you think that the reporting is biased, or you think that the science is biased, or both?

Hank (3:17):

Well, I think that…both. Scientists are people, and so science is going to be biased like people are biased. Reporters are people, and we know that reporters are — I mean, if you just go by — don’t want to get political — but if you just go by political registration, we know that they tilt one way. I would expect that they are going to express themselves — I mean, we know that school teachers are biased, right?

Spencer (3:37):

Well, I think we can be fair, I think it’s fair to say that everyone’s biased, right?

Hank (3:40):

Right, right. So, if you’re in a position where you speak, where it’s your job to communicate something, it’s likely that your bias is going to affect how you communicate and what you communicate. And if you’re the filter between something that’s as arcane and impenetrable as science and a largely scientifically illiterate public — and you’re yourself a scientifically illiterate reporter, probably, right? — what you’re going to do is you’re going to take the most — let’s go back to global warming for a second and then I’ll yield the floor to Cassandra — but you’re going to take the most alarming statement from the IPCC report and you’re going to turn that into a headline and you’re going to ignore the much more likely statement elsewhere in the report where the scientists say, “But that’s not likely to actually happen; it’s more likely to be within this range.” Right? You’re going to — Because that’s what — you know, listen to the report of an earthquake somewhere, and initial reports will always describe an estimate of far more people — almost always, not every single time, but almost always — describe far more people likely killed than actually are in the event. There’s a natural tendency to lead with your strongest headline, and I think that happens in the case of science too. So, we get — we get — we hear that life has been discovered on Mars before we learn, finally, that that little fragment of rock in the Antarctic contains some structures that look a lot like what life produces in some other context; but no, it’s not evidence, necessarily, of life on Mars. Anyway, Cassandra.