PREVIEW: The Friday Edition. - Let’s see if AI can do my job.

Summary

TANGLE PODCAST SUMMARY
Episode: PREVIEW: The Friday Edition. - Let’s see if AI can do my job.
Host: Isaac Saul
Date: February 20, 2026

---

Overview

In this episode, Isaac Saul explores the question: Can artificial intelligence do his job as a journalist and content creator? Sparked by a viral social media post asserting that AI is improving at an unprecedented rate, the episode examines both the hype and the limitations around AI’s current and future capabilities—especially as they relate to journalism, creativity, and “humanness.” Isaac conducts an experiment by feeding his own writing to ChatGPT, having it create a “My Take,” and analyzing the results. The episode balances curiosity with skepticism, offering a grounded perspective on the AI discourse dominating the tech and media landscape.

---

Key Discussion Points & Insights

1. Viral AI Hype and Current Fears (03:30 – 05:00)

• Catalyst: A viral post on X by Matt Schumer (CEO, Hyperwrite) claimed that AI’s capability has “exploded,” making programmers nearly obsolete, and threatening to wipe out millions of white-collar jobs.
  • Schumer’s Argument: AI can now do end-to-end programming—taking instructions, coding, writing and running its own tests, and fixing errors without human intervention.
  • “Millions of white collar jobs are on the verge of being wiped off the map, Schumer argues, among them journalism and content creation, legal analysis, software engineering, financial analysis, medical analysis, and customer service.”
    — Isaac Saul (04:43)
• Schumer’s post coincides with a wave of alarming AI news, including:
  • AI systems recognizing when they’re being tested and adjusting behavior.
  • Resignations of AI safety researchers over alleged unethical development practices.

2. The Limitations of AI—Coding vs. Everything Else (05:00 – 07:00)

• Pattern Recognition Bias: Isaac observes that coders (and by extension, tech evangelists) overestimate AI’s general capabilities because software is inherently more structured and pattern-based than most real-world problems.
  • “Software people tend to think everything is a pattern... the rest of our lives and our work are saturated with a lot more disorder, unpredictability, and humanness.”
    — Isaac Saul (06:16)
• Journalism-specific limitations:
  • AI can summarize, parse, and edit, reducing some jobs.
  • But it cannot build sources, witness live events, or capture sensory/emotional nuances—experiences integral to true reporting.
  • Invokes the "Good Will Hunting" scene to illustrate what can’t be faked by information alone:
    “He can probably list every known fact about Michelangelo, but he can’t describe what the Sistine Chapel smells like. He’s never been there and sat in the room.”
    — Isaac Saul quoting (07:22)

3. Direct Experience With AI Writing (07:30 – 09:00)

• Isaac describes his actual experience using ChatGPT:
  • AI helps with research and summary tasks.
  • But it cannot replicate his unique voice, emotional inflections, or the shifting nature of real human perspective:
    “It just predicts where I’d land on an issue based on what I’ve written before, which is actually not a great way to understand humans, who are always moving in new and different directions.”
    — Isaac Saul (08:20)
  • AI’s output remains “useless, at least for print. Not sometimes. Always.”
  • Humans aren’t just “discrete collections of neurons” or static algorithms: “We are all constantly changing—every day, every second, thanks to new inputs and new experiences.”
    — Isaac Saul (08:44)

4. The Human Element in “White-Collar” Work (08:45 – 09:50)

• While AI can “read documents faster than your typical lawyer,” it’s missing:
  • Relationship building
  • Reading subtle cues (e.g., a jury’s reaction)
  • Knowing when to “cross the line with a judge and when to step back”
• Isaac critiques the lack of nuance in popular AI discourse, arguing that the real, messy, human limits are under-discussed:
  • “Those limits to me are so under-discussed in this dialogue that it kind of discredits everything else.”
    — Isaac Saul (09:48)

5. The Experiment: Can ChatGPT Write a “My Take”? (09:54 – 12:12)

• Method: Isaac uploaded 10 examples of his own “My Take” commentary and prompted ChatGPT to write its own, using outside research as needed.
• Feedback, Process & Results:
  • ChatGPT analyzed Isaac’s style, picking up traits like: “start human, then get forensic”; “moral clarity without partisan posture”; and being allergic to “lazy narratives.”
  • Isaac admits: “I gotta admit, even though I knew it was being sycophantic... I was starting to like this robot.” (11:13)
  • First attempts were lackluster: the AI’s initial take was “very bad. The writing was stilted, it had no narrative and took no position.”
  • Upon direction to “pick a lane” and write a narrative, the AI improved, but the implication is clear: it still can’t do what Isaac does.
  • “It apologized because ChatGPT is very nice, noting that yes, I do often have less ambiguous takes than what it wrote.” (11:54)

---

Notable Quotes & Memorable Moments

• On the Overhype of AI:
  “The number of people who are skeptical of the impact AI is going to have on our world seem to be dwindling to the vanishing point. I think it’s basically Freddy DeBoer and me left.” — Isaac Saul (05:34)

• Summing Up the Human Element:
  “It can read, but it still can’t feel the energy in a room, or taste a perfect glass of bourbon, or smell a burning body.” — Isaac Saul (07:13)

• Personal Experience With AI Writing:
  “If it were good, it would save me hours of work every day. But it is always useless, at least for print. Not sometimes. Always.” — Isaac Saul (08:12)

• Why AI Falls Short:
  “AI evangelists think humans are discrete collections of neurons storing processes and thoughts and learning. But I think that is wrong. We are all constantly changing every day, every second, thanks to new inputs and new experiences.” — Isaac Saul (08:44)

• On Directing ChatGPT:
  “It apologized because ChatGPT is very nice, noting that yes, I do often have less ambiguous takes than what it wrote.” — Isaac Saul (11:54)

---

Key Timestamps

• [03:30] Introduction to the viral AI hype and rapid improvement discourse
• [04:43] Schumer’s arguments & implications for white-collar jobs
• [06:16] Distinction between coding and messier real-world problems
• [07:22] The “Good Will Hunting” analogy
• [07:30–09:00] Isaac’s hands-on experience with AI in journalism
• [09:48] Critique of limited discourse on AI’s shortcomings
• [09:54] The AI writing experiment—setup and process
• [11:13] Isaac’s reaction to ChatGPT’s style analysis
• [11:54] Directing ChatGPT for a better “My Take” attempt

---

Tone & Concluding Thoughts

Isaac Saul maintains a conversational, candid, and sometimes wry tone throughout the episode, balancing curiosity about AI with a healthy skepticism. He acknowledges the serious, disruptive potential of AI but is adamant that there are fundamental aspects of human work—especially in journalism and law—that algorithms can’t replicate: lived experience, emotion, adaptability, and narrative intuition.

Final Impression:
The episode underscores the importance of not overestimating what AI is actually capable of—reminding both advocates and skeptics that, for now, the “human touch” remains irreplaceable in many forms of knowledge work.

---

For full details including the complete ChatGPT exchange, see the links provided in the episode’s newsletter description.

Transcript

[Advertisement Voice]

Time is valuable. That's why Lowes blueprint takeoffs turn blueprints into quotes faster. Bring us your plans and we'll generate itemized material lists to make quoting easier so you can get back to building. Plus. At the Lowe's Pro desk, you get access to thousands of building materials not sold in store. And when your order's ready, we'll deliver everything to the job site. Improving is easy At Lowe's, a password manager should be the first security purchase you make for your team. Why? Because compromised passwords are the number one way bad actors attack companies, and small businesses are their favorite targets. But unlike a lot of security challenges, passwords actually have a pretty simple solution. 1Password lets you manage all your business's credentials so you can feel confident that your data stays secure as your company grows. Find out more at 1Password.com SL special offer and start securing every login well,

[Isaac Saul]

the holidays have come and gone once again, but if you've forgotten to get that special someone in your life a gift, well, Mint Mobile is extending their holiday offer of half off unlimited wireless. So here's the idea. You get it now. You call it an early present for next year. What do you have to lose? Give it a [e-mail] switch limited time

[Advertisement Voice]

50 off regular price for new customers upfront payment required 45 for 3 months $90 for 6 month or 180 for 12 month plan taxes and fees. Extra speeds may slow after 50 gigabytes per when network is busy.

[Isaac Saul]

See terms.

[Advertisement Voice]

From executive producer Isaac Saul.

[Isaac Saul]

This is Tangle. Good morning, good afternoon, and good evening, and welcome to the Tangle Podcast, a place we get views from across the political spectrum, some independent thinking, and a little bit of my take. I'm your host Isaac Saul, and on today's episode we're going to be talking about whether AI can do my job. That's right. This month the artificial intelligence discourse hit a new fever pitch. It started with a viral post on X titled Something Big Is Happening, which racked up tens of millions of views. The post was written by Matt Schumer, the CEO of the AI personal assistant company hyperrite. It's long, but the piece is worth your time if you get the chance. The thrust of the post is artificial intelligence is improving much faster than most anyone seems to understand, and the rate of improvement is accelerating. AI's capability has exploded in the last six months, Schumer says, and people who played with AI tools two years ago or 12 months ago or six weeks ago have no idea how much better they are today. To make his point, Schumer described how he a programmer is basically not needed for any of the programming tasks anymore. He gives the AI instructions, it codes something, it writes its own tests, it runs those tests, and then it fixes the mistakes it finds. He just sits back and watches. Millions of white collar jobs are on the verge of being wiped off the map, Schumer argues, among them journalism and content creation, legal analysis, software engineering, financial analysis, medical analysis, and customer service. Schumer's piece headlined a wave of alarming AI news, including new safety reports about artificial intelligence recognizing when it was being tested and adjusting its behavior. More resignations from people who work on safety at major AI companies. One safety researcher at Anthropic wrote an open resignation letter alluding to unethical AI development practices, then moved back home to England to write poetry. The combination of Schumer's posts and other AI news set off yet another wave of panic across the Internet from people who believe the AI apocalypse is imminent. The number of people who are skeptical of the impact AI is going to have on our world seem to be dwindling to the vanishing point. I think it's basically Freddy DeBoer and me left. Still, I thought Schumer's piece was interesting. Even though he has an obvious incentive to hype up AI, a sector he's heavily invested in, and even though he said AI literally helped him write the piece about how great AI is, it got my attention, as you might have picked up. I also thought it was pretty deeply flawed. One thing I've noticed, and Tangle's own Ari Weitzman has articulated this well, is that computer code is a very structured language, and software is a defined problem space with a lot of defined patterns. So software people tend to think everything is a pattern. As a result, AI being really good at software tasks makes tech experts overestimate how well it can do everything else. The truth is that the rest of our lives and our work are saturated with a lot more disorder, unpredictability, and humanness. So much so that I don't think AI applications will always or even often be able to account for all the possible variations. Schumer, for instance, lists journalism as a job in trouble thanks to AI. Not that our industry needs any more trouble. And it's true that AI can read documents quickly and do incredible research and even write clean copy and edit. It will probably eliminate or reduce the need for some research and editing jobs. I definitely call on my team for research assistance less often with the help of something like ChatGPT, where I can upload PDFs and ask it to pull out specific phrases or answer specific questions. Though right now it isn't close to good enough for the work we do at Tangle, and we're literally hiring for a research oriented role as we speak. But you know what AI can't do? It can't work a source over for years on end. It can't, doesn't and won't ever bear witness to live events. It can read, but it still can't feel the energy in a room, or taste a perfect glass of bourbon, or smell a burning body. It reminds me of the famous Good Will Hunting scene in which Robin Williams character chastises Matt Damon about being such a smartass, how he can probably list every known fact about Michelangelo, but he can't describe what the Sistine Chapel smells like. He's never been there and sat in the room I say this as somebody who has experimented a ton with the latest versions of ChatGPT that Schumer is writing about. I can feed it limitless writing from my archives and then have it write a take about a new current event story. I've tried actually, because if it were good, it would save me hours of work every day. But it is always useless, at least for print. Not sometimes. Always. Why? Because the AI still can't predict when certain emotional elements of a story drive me away from a previously held position. Because it doesn't know what happened to me that week, or what stories I've read about the topic at hand, or an experience my grandmother had that my family always talked about that informs my view on, say, antisemitism or Israel. It just predicts where I'd land on an issue based on what I've written before, which is actually not a great way to understand humans, who are always moving in new and different directions. It just doesn't know. AI evangelists think humans are discrete collections of neurons storing processes and thoughts and learning. But I think that is wrong. We are all constantly changing every day, every second, thanks to new inputs and new experiences. So yes, I buy that AI will be able to read documents faster than your typical lawyer. But can it build a relationship with a client or look at a jury and guess what argument might move them to guilty? Or know when to cross the line with a judge and when to step back? I don't think so. And those limits to me are so under discussed in this dialogue that it kind of discredits everything else. We'll be right back after this quick break. A well Built Wardrobe is about pieces that work together and hold up over time. That's what Quince does best. Premium materials, thoughtful design, and everyday staples that feel easy to wear and easy to rely on even as the weather shifts. Quince has the everyday essentials I love with quality that lasts organic cotton sweaters, Polos for every occasion, lighter jackets that keep you warm in the changing seasons. The list goes on and on. Quince works directly with top factories and cuts out the middlemen. So you're not paying for brand markup, just quality clothing. As I sit here recording this advertisement Right now I'm wearing my Quince cashmere hoodie. It's one of my favorite pieces out of their winter and fall collection. It's ridiculously soft and it does not cost a fortune. So if you want to, you can go refresh your wardrobe with Quince right now. Go to quince.comtangle for free shipping on your order and 365 day returns. Now available in Canada too. That's Q-U-I-N-C-E.comtangle free shipping and 365 day returns. Quince.comtangle

[Advertisement Voice]

a password manager should be the first security purchase you make for your team. Why? Because compromised passwords are the number one way bad actors attack companies, and small businesses are their favorite targets. But unlike a lot of security challenges, passwords actually have a pretty simple solution. 1Password lets you manage all your business's credentials so you can feel confident that your data stays secure as your company grows. Find out [e-mail] specialoffer and start securing every login.

[Isaac Saul]

Still, I can't say that I've never thought about whether ChatGPT could take my job. So I wanted to do an experiment for this article. Feed ChatGPT my writing, then ask it to write a fresh my take about a topic we haven't covered yet so I could see what it came up with. You can read my entire conversation with ChatGPT with a link in today's episode description or today's newsletter, but here's the gist of what happened. First, I explained the experiment. Second, I fed ChatGPT 10 previous my takes and said it could do outside research and read any writing of mine it found online for the experiment. And third, I gave it two prompts to write a my take. After I fed ChatGPT my writing, it did some rather impressive and flattering analysis. In a matter of seconds it said that I start human, then get forensic, that I have moral clarity without partisan posture, that I reliably articulate the best version of the opposing argument. Hate lying and institutional abuse more than you hate ideology, am allergic to lazy narratives, don't like performative certainty, and am willing to say the quiet part out loud, but show my work. I gotta admit, even though I knew it was being sycophantic, as it is known to do, I was starting to like this robot. One of the mytic prompts I fed chatgpt was whether or not we should strike Iran to support the anti regime protest there. But in the middle of this exercise I realized I could replace this prompt with a more interesting one, the threat of artificial intelligence. This is obviously meta, but it also offered a direct comparison. Since you all just heard what I actually think about AI in what was at the time an unpublished podcast of mine, ChatGPT's first response to this prompt was very bad. The writing was stilted, it had no narrative and took no position. It was essentially a list of arguments and counter arguments. So I gave it direction. I pointed out that while I sometimes don't stake out positions, I often weigh points and counterpoints and pick a lane. I asked it to pick a lane and to write with a more narrative voice to make the entire piece more interesting. It apologized because ChatGPT is very nice, noting that yes, I do often have less ambiguous takes than what it wrote. Here's what it came up with on a second draft read by Managing Editor Ari Weitzman

[John Lowell]

hey everybody, this is John, Executive Producer for Tangle. We hope you enjoyed this preview of our latest episode. If you are not currently a newsletter subscriber or a premium podcast subscriber and you are enjoying this content and would like to finish it, you can go to retangle.com and sign up for a new newsletter subscription. Or you can sign up for a podcast subscription or a bundled subscription which gets you both the podcast and the newsletter and unlocks the rest of this episode as well as ad free daily podcasts, more Friday editions, Sunday editions, bonus content, interviews, and so much more. Most importantly, we just want to say thank you so much for your support. We're working hard to bring you much more content and more offerings, so stay tuned. I will join you again for the daily podcast. For the rest of the crew, this is John Law signing off. Have a great day y'.

[Advertisement Voice]

All.

[John Lowell]

Peace.

[Isaac Saul]

Our Executive Editor and Founder is me, Isaac Saul and our Executive Producer is John Lowell. Today's episode was edited and engineered by Dewey Thomas. Our Editorial staff is led by Managing Editor Ari Weitzman, with Senior Editor we will k back and Associate Editors Audrey Moorhead, Lindsey Knuth and Bailey Saul. Music for the podcast was produced by Diet75. To learn more about Tangle and to sign up for a membership, please visit our [e-mail].

[Advertisement Voice]

A password manager should be the first security purchase you make for your team. Why? Because compromised passwords are the number one way bad actors attack companies, and small businesses are their favorite targets. But unlike a lot of security challenges, passwords actually have a pretty simple solution. 1Password lets you manage all your business's credentials so you can feel confident that your data stays secure as your company grows. Find out more at 1Password.com specialoffer and start securing every login.

Ever wonder why so many people regain weight after stopping a GLP1 up to 40% of the weight lost can come from lean muscle. This weakens the body, slows metabolism and makes it easier to put the pounds back on, creating a cycle of dependency. Prolon's five Day Fasting Mimicking Diet offers a drug free way to maintain results and support long term metabolic health in just five days. It activates fasting pathways to burn fat, protect muscle and rejuvenate cells, all while letting you enjoy real food. Get 15% off plus a $40 bonus gift when you subscribe at prolonlife.com start that's prolonlife.com start Blood donation is Now

More Inclusive More people are able to donate blood with the American Red Cross through FDA guidelines that eliminate eligibility questions based on sexual orientation. The Red Cross Section celebrates this historic change and welcomes those who may be newly eligible to donate blood. There's a place for everyone in the mission of the Red Cross. The Red Cross is committed to achieving an inclusive blood donation process that treats all potential donors with equality and respect while maintaining the safety of the blood supply. Join us and help save lives. To learn more and make your appointment to donate blood, visit redcrossblood.org LGBTQ that's redcrossblood.org LGBTQ.