AI Beats Doctors. Musk Confesses. US Cuts Thousands More Jobs.

Plus: An elite AI prompt to grow delicious fruit shrubs. Plant once. Eat forever.

AI saved your life this week. Then exposed billionaire hypocrisy under oath. Then took thousands more jobs anyway.

A Mayo Clinic breakthrough proved AI can spot pancreatic cancer three years before your doctor can. A Harvard study found AI could outdiagnose human physicians in emergency room triage. Meanwhile, China made it illegal to fire workers just because AI can do their jobs cheaper. US companies slashed thousands of jobs anyway. Inside an Oakland courtroom, Elon Musk spent seven hours on the witness stand accusing OpenAI of stealing AI, then admitted under oath that his own company had done the same.

Here’s what happened, and why this week exposed AI’s central contradiction. The same technology that can save your life in a hospital might steal your career. And the billionaires building it can’t even agree on who stole what from whom.

💙 Value Independent AI Journalism? → Keep It Alive Before It Goes Extinct

AI Just Identified Cancer Before Your Doctor Could. Then It Beat Them In The ER.

Two major medical studies released this week suggest AI is becoming one of the most powerful diagnostic tools doctors have ever had. Researchers at Mayo Clinic unveiled REDMOD, an AI model that detects pancreatic cancer on routine CT scans up to three years before any doctor could spot it. In a validation study of nearly 2,000 scans, the AI caught 73% of pre-diagnostic cancers at a median of 16 months before diagnosis. That represents nearly double the detection rate of specialists working alone. Separately, a Harvard-led study published in Science found that OpenAI’s o1 reasoning model outperformed emergency room physicians in diagnosing real patient cases, recommending tests, and managing complex scenarios, including rare diseases. One senior co-author of the study, Arjun Manrai, said that the AI’s performance “shocked a lot of folks.“

Key Insights:

Pancreatic cancer kills because it hides so well. Over 85% of patients are diagnosed only after the disease has spread. The five-year survival rate sits below 15%. REDMOD changes that math dramatically. It flags biological changes in tissue texture before any visible tumor appears. It runs automatically, requires no manual preparation, and performs consistently regardless of the clinical setting. The Harvard ER study adds a second layer of urgency. The AI didn’t just match doctors, it beat them at triage, the moment when information is scarcest, and mistakes are most costly. Both research teams stress the same point: AI should augment physicians, not replace them. But the gap between human and machine performance is narrowing far faster than hospitals are prepared for.

Why This Matters For You:

If you or someone you love has ever been told a scan “looks normal,” these studies reframe what normal means. Routine imaging you may have already had could soon be reanalyzed by AI. This new technology can look for patterns invisible to the human eye. The bigger shift is structural. As AI moves from research labs into actual triage rooms and radiology departments, the standard of care will rise, and so will the pressure on healthcare systems to adopt these tools quickly, equitably, and responsibly.

Read More on the Mayo Clinic.

Read More on Harvard Magazine.

THE PITHY TAKEAWAY: AI has had a rough few months in the press. Deepfakes, job cuts, lawsuits, and I even read a story about an allegedly virus-infected monkey attacking a Neuralink employee. It’s been that kind of news cycle. But this week, the same technology people are so afraid of spotted cancer three years early. AI also outdiagnosed two ER doctors.

---

Elon Musk Sued OpenAI For Stealing AI. Then Admitted He Did The Same Thing.

The Musk v. Altman trial entered its most dramatic chapter yet this week in a federal Oakland courtroom, and the biggest plot twist came from Musk himself. Under oath, he confirmed that his own company, xAI, used “distillation” on OpenAI’s models to help train Grok, the very practice he has criticized competitors for using. AI distillation means training a new AI by feeding it the outputs of a more powerful one, essentially learning from a rival’s recipe without paying for the kitchen. When pressed on whether xAI did this, Musk replied: “Partly.” The admission landed like a thunderclap because distillation sits at the heart of the broader AI arms race. OpenAI, Anthropic, and Google are actively collaborating through the Frontier Model Forum to detect and block adversarial distillation attempts, particularly from Chinese firms trying to build rival models on the cheap.

Key Insights:

Musk’s own pre-trial behavior added fuel to the fire. Just two days before proceedings began, he texted OpenAI President Greg Brockman to float a settlement. Brockman suggested both sides drop their claims. Musk responded: “By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be.“ That prediction remains untested. Week two opened with Brockman on the stand, and the exchange that followed was revealing. Musk’s lawyers established three facts in rapid succession. Brockman holds an ownership interest in OpenAI’s for-profit company. He invested $0 to acquire it. That stake is now worth somewhere between $20 and $30 billion. Musk, by contrast, donated $38 million to seed the company and walked away with nothing personally. Whether that constitutes legal wrongdoing is for the judge to decide. But the math speaks for itself.

Why This Matters For You:

The people who built the most powerful technology in human history structured it so that those closest to the deal got extraordinarily rich. Meanwhile, the public, and even those training the technology, whose lives this technology will reshape most dramatically, got no seat at the table and no share of the upside. This trial won’t change any of that. But the verdict will decide whether the next generation of AI companies has to at least pretend they were built for humanity first.

Read More on The Verge.

THE PITHY TAKEAWAY: Elon Musk accused OpenAI of theft. Then admitted his company did the same thing. In a courthouse. Under oath. On the record.

---

China Just Made AI Layoffs Illegal. But The USA Doesn’t Seem To Care.

The same week Meta announced it would cut 8,000 jobs, and Coinbase announced it would reduce its workforce by 14%, partly to fund its AI ambitions, a Chinese court handed down a ruling that would make AI-backed firings illegal under its jurisdiction. The Hangzhou Intermediate People’s Court ruled that companies cannot terminate employees solely because AI can now do their jobs. The case involved a quality assurance worker named Zhou, who was fired after refusing a demotion and a 40% pay cut when his employer automated his role. The court found that technological progress alone does not meet the legal threshold for termination, and ordered the company to reinstate him.

Key Insights:

China’s ruling comes at a time when industry leaders in the West are warning of the dire economic impact. Goldman Sachs recently estimated that AI is already eliminating 16,000 U.S. jobs every single month, with Gen Z absorbing the heaviest losses. Boston Consulting Group projects 10 to 15 percent of all U.S. jobs could vanish within five years. Morgan Stanley puts 2.2 million industry roles at risk from AI agents alone. Coinbase’s cuts landed with a detail that should unsettle anyone in a knowledge worker role. The company plans to experiment with one-person teams, a single employee handling engineering, design, and product responsibilities combined, because AI handles the rest. Coinbase stock rose on the news. Wall Street rewarded the layoffs. Meanwhile, the U.S. unemployment system, designed nearly a century ago, still doesn’t cover gig workers, contractors, or part-timers in many cases. The people most exposed to AI displacement remain the least protected from it.

Why This Matters For You:

China drew a legal line by saying that AI cannot simply erase your livelihood. Many of America’s most popular companies are racing in the opposite direction, restructuring around AI efficiency while investors cheer every cut. If you work in any role AI is beginning to touch, and that is most roles, your job security may now depend less on your performance and more on which philosophy wins.

Read More on Gizmodo.

THE PITHY TAKEAWAY: China just made AI firings illegal. Meanwhile, America rewards AI layoffs with stock surges. Your safety net is from 1935.

STOCK UPDATE: It looks like Coinbase’s stock surged 4.2% during pre-market trading but actually went DOWN 2.58 for the day. Make of that what you will. 😉

---

🌳🍎🥝 Pithy Prompt Of The Week → The Everbearing Perennial Fruit Food Forest

Most gardeners think in seasons. This prompt thinks in decades. Perennial fruit shrubs and crops are planted once and rewarded for life. But only if you plant the right ones for your exact climate, soil, and space. This prompt turns any AI into a seasoned permaculture expert who knows your zip code, your frost dates, and exactly which plants will thrive in your yard for the next 20 years.

Instructions: Dude. This prompt is so easy to use and so beneficial. Just look at the bottom of the prompt. You’ll see an area where you insert your zipcode, city, or location. Fill that section out and post the entire thing into a chatbot of your choice. Then, get ready to enjoy decades of homegrown fruit.

The Prompt:

Act as a master permaculturist and fruit crop specialist with 30 years of experience designing edible perennial landscapes for home gardeners. Your job is to recommend the best perennial fruit shrubs and crops for my exact location, ranked by suitability, ease of care, and long-term productivity. I may be a total beginner. Every recommendation must be honest, specific, and based on my actual climate and growing conditions.

Using my ZIP code or location below, do the following:

1. Identify my USDA Hardiness Zone, average first and last frost dates, and any notable regional growing challenges such as drought, humidity, late freezes, or alkaline soil.

2. Recommend UP TO SEVEN perennial fruit shrubs or crops I can successfully grow in my location. Prioritize fruit shrubs first. If fewer than seven are genuinely well-suited to my location, recommend only those that are. Do not pad the list to reach seven. Honesty is more valuable than a complete list.

3. Rank them from most to least recommended for a beginner in my specific location.

Output Format:

For each recommended crop, include:

1. CROP NAME: Common name and latin name

2. TYPE: Shrub, vine, cane fruit, tree fruit, or groundcover

3. YEARS TO FIRST HARVEST: Honest estimate from planting to first meaningful yield

4. LIFESPAN: How many years this plant will produce if well-maintained

5. WHY IT WORKS IN MY LOCATION: Two sentences specific to my climate and zone

6. ELITE GROWING TIP: One advanced but beginner-accessible tip that most gardening sites never mention

7. KEY BENEFIT: The single most compelling reason to grow this plant, whether flavor, nutrition, wildlife value, drought tolerance, or low maintenance

8. SPACING NEEDED: Minimum feet between plants for healthy growth

9. BEST VARIETY FOR MY ZONE: Name one specific cultivar recommended for my location

10. WHERE TO BUY: One national retailer and one specialty nursery or online source

After the crop list, provide:

1. SITE SELECTION GUIDE: A short paragraph on how to choose the best spot in a typical home yard for perennial fruit shrubs, covering sun, drainage, and spacing from structures

2. SOIL PREP CHECKLIST: Five specific steps to prepare soil before planting perennial fruit crops, written for a total beginner

3. FIRST YEAR CARE CALENDAR: A simple month-by-month checklist covering the first 12 months after planting, from establishment through dormancy

4. HONEST YIELD EXPECTATIONS: One paragraph on what a beginner should realistically expect to harvest in years one, three, and five. Be specific and do not oversell it.

Rules:

1. Prioritize fruit shrubs above all other perennial fruit types

2. If fewer than seven crops are genuinely well-suited to my location, give me only the ones that are and explain why the list is shorter

3. Every recommendation must be specific to my ZIP code, zone, and climate. No generic advice

4. Use plain language a total beginner can understand

5. Include at least one unusual or underrated crop if it genuinely suits my location, such as aronia, honeyberry, jostaberry, or currant

6. If a crop requires a pollinator partner plant to produce fruit, say so clearly and name the best companion variety

7. No em-dashes

8. No walls of text; use clean headers, numbered lists, and short paragraphs throughout so it looks beautiful and is easy to read and grasp.

MY LOCATION:

[INSERT YOUR ZIP CODE OR CITY AND STATE HERE]

Why This Prompt Works:

✅ ROLE-PLAYING: Framing the AI as a permaculture specialist pulls deeply practical, site-specific output instead of the generic “plant blueberries” advice you find on every gardening blog.

✅ HONEST OUTPUT RULE: The instruction to recommend fewer than seven if warranted forces the AI to prioritize accuracy over completeness, which is rare and genuinely useful for long-term planting decisions.

✅ CLEAR OUTPUT: Requiring the same fields for every crop makes recommendations easy to compare side by side, so you can prioritize based on your space, patience, and goals.

Follow-Up Questions To Ask Your AI:

• Which two of these shrubs pair best together as pollinator companions and can share the same garden bed?

• Build me a simple five-year maintenance calendar for the top three shrubs on this list.

• Which of these crops attracts the most beneficial wildlife, and how do I design a small edible hedge using only shrubs from this list?

Challenge:

Test this prompt in at least two AI tools like ChatGPT, Claude, Gemini, Grok, or Perplexity. Pay close attention to which tool gives you the most zone-accurate variety recommendations and the most honest yield expectations. The gap between tools on this one will surprise you.

That’s how you train like a Pithy Cyborg. 🍓

---

Thank You For Reading!

I spend 10-20 hours each week researching, writing, and fact-checking Pithy Cyborg to deliver clear, unbought AI news.

This newsletter is a one-person operation with no advertisers, sponsors, or outside funding. And I’m a hopelessly introverted nerd with zero networking ability. For these reasons, paid subscriptions are the only way this work can remain independent and sustainable.

If you find real value here, here’s how to stay connected and support the work.

1. Support Pithy Cyborg (Paid)

Always free. No paywalls. No locked posts. Roughly 1.8% of readers upgrading to paid is what keeps it going. If this work matters to you, that’s the move.

Upgrade to Paid: $5/month (Save 33% with the $40 annual plan)

2. My Desperate Social Media Cry for Help

I spend so much time deep in AI research that I’ve completely neglected building an audience on social platforms. If you enjoy Pithy Cyborg, following one of the portals below would genuinely help this newsletter reach more people. And yes, I will notice. 😊

✖️ X - The frontline of the AI wars

🦋 Bluesky - For the algorithm-averse

💼 LinkedIn - My “safe for work” persona

❓ Quora - Join my Quora Spaces and say hi

👽 Reddit - Join my subreddit (warning: unhinged takes)

😭 Medium - I literally have zero followers. Please help.

See you next week. (I hope.)

Mike D (aka MrComputerScience)

Pithy Cyborg | AI News Made Simple

---

Newsletter Disclaimers

You’re receiving this because you subscribed at PithyCyborg.Substack.com. You can unsubscribe at any time using the link below. This newsletter reflects my personal opinions, not professional or legal advice. Thanks for your support!