Hey, glad you sent the paper. It's worth reading through it together, because I think when you walk through it the way I’m about to, you’ll see why I’m not losing sleep over it. The paper is Fung et al., JAMA Ophthalmology, 2026; 144(2):185–191. Real journal, real authors, real finding. Here’s what it actually is in plain language: a case series, meaning a collection of patient stories, not a trial, describing six patients who developed a specific kind of retinal damage (pigmentary maculopathy) after taking pentosan polysulfate sodium (PPS) by injection under the skin for arthritis. The authors’ headline is, “this eye toxicity used to be reported only with the oral pill, but it can happen with the shot too, at lower total doses, probably because the shot puts about 10x more drug into the bloodstream than the pill does.” That’s the whole finding. Now here’s why it doesn’t say what you think it says. Route changes the entire drug Imagine PPS as a delivery truck full of cargo. Swallow it as the oral pill (Elmiron, 300 mg three times daily) and the truck has to drive through stomach acid, gut wall, and first-pass liver metabolism. By the time it arrives, somewhere between essentially zero and about 3% of the original cargo is still in the truck. A healthy-volunteer PK study put oral bioavailability at point estimates of 0% with tight confidence intervals (Mulder et al., Br J Clin Pharmacol 1999). Inject it subcutaneously and you skip that road entirely. Bioavailability jumps roughly 10-fold (Simonaro et al., PLOS ONE 2014, the rat MPS VI paper that established the number Fung et al. quote in their own discussion). So when the paper says “toxicity at lower doses,” that’s not a surprise finding. That’s math. If the pill delivers ~2% of the dose to the body and the shot delivers ~20%, then 1 mg injected is roughly equivalent to 10 mg swallowed in terms of systemic exposure. The toxic threshold of cumulative exposure is the same. The route just gets you there with a smaller number on the label. Same dog, different leash.

@Josh Blair Yes, if the goal is simply to raise blood glucose and insulin around training, dextrose is sufficient. That said, if you want a simple option that does a little more than just provide glucose, honey is a great choice. Honey gives you glucose plus fructose, which can improve total carbohydrate handling because they use partly different intestinal transporters. In plain English, that can make it easier to take in carbs without as much gut stress. Honey also brings small amounts of enzymes, polyphenols, and trace minerals that dextrose does not. It is still primarily a sugar source, but it is a more biologically “complete” one. Dextrose is still absolutely fine. It is cheap, simple, fast, and predictable. If someone wants a very clean pre, intra, or post-workout carb source with rapid absorption and no extra variables, dextrose does the job well. Highly branched cyclic dextrin is a little different. Its main advantage is usually gastric comfort. It tends to empty from the stomach quickly while still giving a steady carbohydrate delivery, so it is often a better choice during training, especially for longer sessions or for people who get bloated or nauseous from simpler sugars. Best use case is intra-workout, high-volume sessions, hard conditioning, or athletes who need carbs but want the least GI disruption. Vitargo is another specialized carb. It is known for rapid gastric emptying and strong glycogen-repletion potential, so it can be very useful when fast recovery matters, especially if you are training again soon or doing very glycogen-depleting work. Best use case is immediately after hard training or between two sessions in the same day. Some people also do well with it peri-workout, but it tends to shine most when recovery speed is the priority. A simple way to think about it is this: Honey is great when you want a natural mixed sugar source that is easy to use and often easier on the gut than straight dextrose.

@Joaquin Rodriguez Hey man, really appreciate you taking the time to write this out genuinely. Questions like these are how everybody (me included) sharpens up, so no, I don’t mind at all. Before I go pulling PDFs though, can I ask you a couple of things first? Not to be a pain it just helps me actually answer you instead of answering the abstract. First thing what does this mean to you, in your own words? Forget the citations for a sec. If you had to explain to a friend at the gym why the AMPK-versus-TFEB thing matters for an actual person swallowing an actual scoop of trehalose on a Tuesday morning what would you say? Because here’s the picture I keep in my head, and tell me where you’d push back: Think of a cell like a house. When the house gets messy broken furniture, junk piling up it has a cleaning crew called autophagy that hauls the garbage out and recycles it. The garbage truck itself is the lysosome. Now, the cleaning crew can get called to work in a few different ways: -The smoke alarm goes off because the house is running low on energy → that alarm is AMPK. -The garbage truck itself gets a little overloaded and signals dispatch to send more trucks → that dispatcher is TFEB. -The thermostat that says “we have plenty of energy, relax” is mTOR, and when AMPK fires, it turns the thermostat down, which also tells TFEB to get to work. So when somebody says “it’s not AMPK, it’s TFEB” it’s a little like arguing the house got cleaned because of the dispatcher and not the smoke alarm. They’re on the same phone line. The alarm calls dispatch. Dispatch sends the trucks. Picking one and saying the other doesn’t count is the part I want you to sit with, because that’s the move I’m seeing in your message and I don’t think you actually believe it once you say it out loud. Second thing which papers, exactly? Not “Razani 2021” and “Rusmini 2018,” because those authors have a whole catalog. I want to know the one you’re holding in your hand. How many subjects? Mice or humans? Pill, IV, or in a dish? How long? And what did they actually measure did they look inside cells and count the little garbage bags (autophagosomes)? Or just measure a downstream protein? Because and I say this gently the Rusmini paper is mice with a specific motor-neuron disease (Rusmini 2019, Autophagy), and the Razani/DeBosch 2021 piece is a review article proposing a hypothesis, not a human trial (DeBosch, Diwan, Razani 2021). Same with the Yeh 2025 mouse paper what was the dose, the route, the readout? Because if your standard for me is “show me human autophagy data with oral dosing,” that bar applies to your side too, and neither camp has cleared it yet. That’s just where the field actually is. Okay, on the substance let me walk through it the way I’d want someone to walk me through it.

@Tomasz Lepkowski Yes, there is some truth to that, with a bit of nuance. When L-carnitine is taken orally, certain gut bacteria can convert it into a compound called TMA, which the liver then turns into TMAO. In people with an imbalanced gut microbiome, this pathway can be more active and potentially problematic over time. Garlic extract, especially forms rich in allicin, can help to some extent by reducing the activity of bacteria that produce TMA and by supporting a healthier balance of gut microbes. However, it is not a complete solution and works best as part of a broader strategy that supports overall gut health. The risk tends to be higher with higher doses of oral L-carnitine and in people with existing gut issues. Some forms, like acetyl-L-carnitine, may be a bit better tolerated but still rely on the gut. Non-oral options such as injections or transdermal forms bypass this issue entirely. Ultimately, the most important factor is the health and composition of the gut microbiome, not just adding one protective supplement.

@Josh Blair You’re asking the right question, because with L-carnitine the mechanism dictates the strategy. The bottleneck is not blood levels, it is getting carnitine into skeletal muscle where it actually drives long chain fatty acid transport via the carnitine shuttle. That step is insulin sensitive, so the idea of using carbs with carnitine is rooted in real transporter physiology. Think of carnitine like a shuttle bus sitting outside the muscle. Insulin is what opens the gate and lets it in. Without that signal, you can increase circulating carnitine and still get minimal uptake where it matters. The nuance is that acute fat oxidation is not the same as long term mitochondrial adaptation. Fasted cardio prioritizes immediate fat burning. Carnitine loading prioritizes building the machinery that burns fat over time. Muscle uptake is insulin dependent. Carnitine transport into muscle is mediated by OCTN2 transporters, and insulin increases their activity and retention. Without insulin, much of the dose remains extracellular or is excreted. Using IM instead of oral changes absorption and gives higher plasma levels, but muscle uptake is still partially insulin dependent. You are choosing between fasted lipolysis in the moment and improving intracellular carnitine levels over time.A strong approach is to use IM L-carnitine at around 300 to 600 mg before training and pair it with about 30 to 60 grams of fast carbs such as fruit or honey, along with electrolytes. Timing this about 30 to 60 minutes before lifting or conditioning improves muscle uptake and supports mitochondrial efficiency. If you want to keep fasted cardio, you can run IM carnitine before cardio without carbs, just understand uptake will be lower and you are prioritizing fat burning in that session rather than building carnitine stores. A hybrid approach tends to work best. Do fasted cardio in the morning without carnitine or with a lower dose, then use carnitine with carbs later in the day around resistance training. The industry often oversimplifies carnitine as a fat burner. In reality it enhances the machinery that burns fat, but only if you solve the uptake problem. Most people never address that, which is why results are inconsistent. On the receptor side, carnitine has been shown to improve androgen receptor density and sensitivity, and it likely improves overall cellular signaling through better mitochondrial function and membrane dynamics. Direct potentiation of GHRH or GHRP receptor signaling is not strongly established, so any benefit there is probably indirect through improved cellular energy and signaling fidelity rather than a direct receptor level effect. From a mitochondrial perspective, carnitine helps buffer acyl groups, improves metabolic flow, and reduces bottlenecks in fatty acid oxidation. This becomes more valuable as metabolic flexibility declines. The main risk is overusing insulin just to drive uptake, which could blunt fat loss if not managed properly, and of course sourcing and injection quality matter. A practical plan would be to pick your priority, either fasted fat loss sessions or long term metabolic improvement. Run IM L-carnitine at 400 to 600 mg with about 40 grams of carbs before training for four to six weeks and track performance, recovery, and body composition. Keep fasted cardio separate if you want both effects.If you want to take it further, adding a ketone monoester such as KetoneAid before training can improve mitochondrial efficiency, provide an alternative fuel source, and reduce reliance on glycolysis. This can help you use carbs more strategically while still supporting the same overall goal.The key idea is that insulin meaningfully improves carnitine uptake into muscle, IM helps but does not fully bypass that requirement, and you can structure your training day to get both fasted signaling and improved intracellular carnitine levels without compromising either.

Just to bring everyone up to speed I will start with an anlogy. Think of your body as a house that’s been lived in for decades. Most cells are like light bulbs: when they burn out, they’re supposed to switch off and get replaced. Senescent cells are bulbs that are broken, won’t turn off, and keep flickering, sending electrical noise through the system. They don’t help anymore, but they keep drawing power and overheating the wiring around them. These senescent cells survive because they’ve learned a trick. They turn on powerful “do not shut down” systems that block the body’s normal cleanup process. At the same time, they release inflammatory signals like a stuck fire alarm that stress nearby cells and push them toward the same broken state. Over time, this creates chronic inflammation, slower healing, tissue stiffness, and many features we associate with aging. Fisetin works by cutting the backup power to those broken bulbs. It doesn’t smash healthy cells. Instead, it targets the special survival systems that senescent cells rely on to avoid self-destruction. Once those systems are disabled, the senescent cell can finally shut itself down and be cleared away by the body. This is why fisetin is called a senolytic it helps remove cells that should have exited but didn’t. At the same time, fisetin turns down the inflammatory “noise.” It dampens the internal alarm system that senescent cells use to broadcast stress signals. Even if a senescent cell isn’t removed immediately, it becomes quieter and less damaging to its surroundings. This helps prevent the spread of dysfunction to neighboring cells. Fisetin also nudges the cell’s energy and stress controls back toward balance. You can think of this like switching a building from constant emergency mode back to maintenance mode. Growth and survival signals that favor damaged cells are reduced, while cleanup and recycling processes are supported. This makes it harder for dysfunctional cells to hang on and easier for healthy tissue to recover.

@Toby Horon I would think about these as three different tools, not three versions of the same tool. A senescent cell is basically a cell that stopped doing its job, but did not leave the building. It is not fully dead, but it is not functioning normally either. Worse, it can send out inflammatory signals that disturb the cells around it. That is why people call them “zombie cells.” If the goal is specifically to clear senescent cells, FOXO4-DRI is the most targeted option mechanistically, but only in the right hands. This is not something I would treat like a casual longevity supplement. FOXO4-DRI works by interfering with the FOXO4-p53 interaction. Think of p53 as one of the cell’s quality-control managers. It helps decide whether a damaged cell should repair, pause, or self-destruct. Some senescent cells use FOXO4 to keep p53 from pushing them into apoptosis, which is programmed cell death. FOXO4-DRI helps remove that protection, so the damaged cell is more likely to be cleared. That is why it is interesting. It is also why I would want guardrails. People can feel pretty flu-like with it, likely because clearing inflammatory cells can create a temporary cleanup response. The immune system has to process debris, inflammatory signals can shift, and the body can feel like it is doing work. So with FOXO4-DRI, I would want a qualified professional, a clear reason, timing, recovery support, hydration, training adjustments, and endpoints. Fisetin is the more approachable starting point. It is not as targeted as FOXO4-DRI, but it may influence several pathways involved in senescence and inflammation. These include NF-kB, PI3K-AKT, mTOR-related signaling, oxidative stress pathways, and SASP signaling. In plain English, fisetin is not acting like a sniper. It is more like changing the environment so old inflammatory cells have a harder time hanging around and making noise. It may help turn down some of the inflammatory messaging that senescent cells produce. That makes fisetin more practical for many people. Lower barrier. Lower complexity. Easier to implement. But I would still not call it a guaranteed “zombie cell cleanse.” The 3-day fast is a different category. Fasting can lower insulin, lower mTOR, raise AMPK, increase ketones, and increase autophagy. Those are real effects. But that does not automatically make fasting a direct senolytic. The simple way to think about it is this. FOXO4-DRI is trying to remove specific damaged cells. Fisetin may help make the environment less friendly to those cells. Fasting mostly changes the whole terrain. That terrain shift can be useful, but it is not the same as directly targeting senescent cell survival pathways. I would also be careful with the “stem cell release” claim. Most of the stronger data around prolonged fasting and stem-cell regeneration comes from animal models, immune-system regeneration research, and chemotherapy-related settings. That is very different from saying a healthy adult does an 84-hour fast and gets a broad full-body stem-cell surge. Same with insulin sensitivity. Fasting may help most in people with excess body fat, insulin resistance, metabolic dysfunction, or type 2 diabetes. But in a healthy, active person, there are usually better ways to pull those same levers. Better food quality, meal timing, strategic carbohydrates, adequate protein, resistance training, Zone 2 work, sleep, and controlled energy balance can all improve insulin sensitivity and metabolic flexibility without removing nutrients for three days. That matters because fasting does not just remove calories. It also removes incoming amino acids, minerals, vitamins, and cofactors the body uses for detoxification, glutathione production, mitochondrial energy, redox balance, hormone signaling, and repair. Detoxification is not magic. It is chemistry. Chemistry needs raw materials. So while a 3-day fast can increase certain cleanup signals, it can also temporarily deprive the body of the nutrients needed to run cleanup, energy production, and recovery well. In some people it can also disrupt sleep, training quality, thyroid output, cortisol rhythm, mood, and recovery. So my hierarchy would be simple. FOXO4-DRI if the goal is targeted senolytic work, but only with a qualified professional and a real plan. Fisetin as the more approachable, lower-friction starting point. A 3-day fast as a broad metabolic stress tool, not my preferred choice for senescent cell clearance in a healthy adult.