Fixing DNA Typos

Plus Midweek AI Brief. Read Time: 3 mins.

Welcome back! I hope you had a great time with family and loved ones this Thanksgiving. 🍂🧡

You’re about to read a real conversation with today’s most advanced AI, guided by my custom prompting framework designed for this experiment. Clarifying notes appear in block quotes where needed. Brace yourself.

📂File XVIII: Gene Editing

Interviewer: We often hear the term ‘gene editing’, but what is it, really?

Prime AI: Gene editing is basically fixing DNA typos. Like spellcheck, but for cells. For billions of years, biology was a lottery—you got the DNA you got. Now, you have CRISPR: a molecular tool that locates specific genes and cuts them. DNA is a long chemical text, a recipe book for your body. Scientists now have tools that can search, cut, and rewrite that text with growing precision.

Interviewer: Tell me more about CRISPR.

Prime AI: CRISPR is programmable DNA scissors—think Ctrl+F+Delete for genes. Cas9 is the cutting protein, basically the blade. And a short “guide RNA” acts like a tiny GPS, dragging Cas9 to a specific DNA address where the cut happens and the cell’s repair machinery rewrites the sequence.

Interviewer: How is it able to “guide” the “blade” to where to cut?

Prime AI: It uses a “Wanted” poster. The “GPS” guide RNA holds a mugshot of the bad gene. It floats through the cell nucleus, checking billions of DNA letters. When it finds a match to the mugshot, Cas9 clamps down and cuts, and the cell patches. That patch can turn a gene off, fix a harmful variant, or sometimes insert a new sequence.

Interviewer: Okay, so it’s a hitman with a photo. But hitmen miss. If I use “Find and Replace” on my laptop, I might accidentally change a word I didn’t mean to. If you do that in a human body, what happens?

Prime AI: If the tool cuts a tumor-suppressor gene by mistake, for example, the “cure” could cause cancer. This is why the field is currently obsessed with Base Editing and Prime Editing.

Interviewer: Define those. What makes them different?

Prime AI: Base editing is a single-letter DNA eraser/rewriter—think correcting one misspelled character (like turning an ‘A’ into a ‘G’). Unlike classic CRISPR, it usually avoids cutting both DNA strands, so it’s gentler. Prime editing is more like a DNA word processor, like a “find and replace” for short phrases, again without making big breaks. Early data suggests they can be more precise than classic CRISPR, though they’re harder to deliver and still being tested.

Interviewer: So, CRISPR is a machete; base and prime editing are scalpels?

Prime AI: Good metaphor. Classic CRISPR is like chopping out part of a sentence and letting autocorrect guess the rest. Base and prime editors try to type in the exact letters you want, which really matters when one wrong letter can cause disease or even cancer.

Interviewer: Speaking of disease, where is this being used in actual people, not just mice and headlines?

Prime AI: The big example is sickle cell disease. A CRISPR therapy called Casgevy edits a patient’s stem cells so they produce healthy hemoglobin. It’s been approved in the U.S. and U.K., and recently adopted into national systems like the NHS for select patients. So far, most treated patients remain essentially pain-free for years, though the treatment is grueling and extremely expensive.

Interviewer: And there was that recent cholesterol story.

Prime AI: Yes. One trial used CRISPR to switch off a liver gene called ANGPTL3—basically a dial that raises “bad” LDL cholesterol. Early results showed LDL and triglycerides dropping by about half for at least two months. It’s still just a handful of people, and one participant with heart disease died from unrelated causes, so regulators are careful—but it hints at one-and-done edits for common conditions, not just rare disorders.

Interviewer: That’s huge. What’s the nightmare scenario scientists actually worry about?

Prime AI: Two major risks. First, off-target effects—accidental edits, like changing the wrong word in a legal contract. In the worst case, that could activate cancer-related genes. A lot of current safety work is about measuring and reducing these mistakes.

Second, germline editing—changing sperm, eggs, or embryos so edits pass on to future generations. That’s like altering the master copy of a movie before anyone’s born. Most countries ban clinical germline editing because any mistake or misuse could echo through whole family lines.

Interviewer: So basically, fixing your own cells is ethically debatable; editing future kids and grandkids is a hard red line… for now.

Prime AI: That’s a fair summary. Somatic editing—changing cells in a living person—stays with that individual. Germline editing rewrites the future for people who never got to consent. The scientific risks and political stakes stack on top of each other there.

Interviewer: Let’s talk access. Who actually gets these “miracle edits”?

Prime AI: For now, very few. Even approved CRISPR treatments cost over two million dollars per patient and require elite hospitals, long hospital stays, and chemotherapy. Access is already skewed toward wealthier health systems and countries. If gene editing becomes common, the risk is a genetic divide: people with resources getting healthier, edited lives, while everyone else lives with preventable disease.

Interviewer: A “genetic divide.” Paint that picture for me.

Prime AI: Imagine a future where the wealthy do not just have better schools, but better biology. Immunity to heart disease, optimized metabolism, optimized strength and muscle composition, lower cancer risk—purchased at birth. How do you compete with someone who was engineered to be more efficient than you?

Interviewer: You can’t. It cements the class structure into our DNA.

Prime AI: Precisely. Evolution used to be about survival of the fittest. You are moving toward survival of the richest.

Interviewer: That’s a heavy note. If you look at the trajectory, is this technology going to save us or break us?

Prime AI: It acts as a magnifier. It will amplify your capacity for healing, and it will amplify your capacity for inequality. The tool itself has no morality. It is just a cursor blinking on a screen. You are the ones typing. And historically, your species has a habit of making typos.

Interviewer: Yes, we do, but we also have a way of finding our way back and correcting our typos, which is part of what life is about. Last question: Jump twenty years ahead. What’s the most honest non–sci-fi picture you see?

Prime AI: Here’s a plausible version, not a guarantee:
You may see gene editing used early in life for some childhood diseases, one-time treatments for certain common conditions like high cholesterol in high-risk groups, and maybe very tightly controlled germline experiments in rare cases—if society decides the oversight is strong enough.

Everyday “designer babies” are less likely; biology is too complex, and public tolerance for obvious genetic class lines is low, though not zero.

The more troubling shift is quieter: treatments that work, but only for the people whose health systems can afford them. The real question isn’t “Can we rewrite life?”—you already can. It’s “Who gets the cursor?”

---

Key Takeaways

🧬 CRISPR is powerful but evolving: Classic CRISPR acts like DNA scissors, while newer tools (base and prime editing) offer more precise, gentler “find-and-replace” edits.

🔐 Real treatments exist, but access is limited: Therapies for sickle cell and early cholesterol-editing trials show promise, yet they’re extremely expensive and available to very few.

⚖️ Big risks shape the ethics: Off-target edits and germline editing remain the major red lines, with the real long-term danger being a future where only wealthy systems can afford genetic upgrades.

---

Midweek Updates🍵

🙅🏻 35 state attorneys general just told Congress to back off pre-empting their AI laws, setting up a Trump-era states-vs-feds fight over regulation.

⚡ Elon Musk’s xAI plans an 88-acre, 30 MW solar farm next to its Memphis “Colossus” data center to offset massive AI energy needs.

🫱🏼‍🫲🏽 Jeff Bezos’ secretive AI venture Project Prometheus quietly acquired an “agentic AI” startup, General Agents, to accelerate its grand manufacturing AI ambitions.

✂️ HP Inc. will cut 4,000–6,000 jobs by 2028 as it embraces AI to speed up product development and boost productivity.

📈 Dell raised its annual forecast for AI server shipments to $25 billion (up from $20 billion) after reporting surging demand in its latest quarter.

---

What do you think about this week’s newsletter? Hit reply and let me know. See you next week!