Molnupiravir is special in that it works by making viral replication more error-prone, with so many mutations that the virus cannot reproduce at all. If a virus manages to replicate anyway, it may do so with more mutations than normal, and this could theoretically increase the chance of a new variant arising.
Paxlovid's mechanism doesn't involve inducing mutations, so it doesn't have this concern. It's a protease inhibitor, blocking the virus's ability to break down proteins. Viruses can still develop resistance to this class of drugs, but only using their natural mutation rate.
Doesn't seem that bad to me, the virus doesn't need Molnupiravir to mutate. It can do very well by itself, in the billion people and animals it infected. The fact we are seeing convergent evolution shows that it is naturally very good at this.
By increasing the mutation rate way beyond the sweet spot (that what's Molnupiravir does) the virus may get some desirable mutations, but won't be able to hold on to them. And if, thanks to the drug, the patient is cured, it is bad news for the virus, those who are not infected can't produce new variants.
But like antibiotics, I think this is an "all or nothing" drug. Either give maximum dose to do as much damage as possible, or don't give anything at all. Too small a dose may promote building resistance instead.
It was 2020 when the drug was undergoing development.. literally throw everything at the wall and try to stave off the mass death from uncontrolled spread. Fortunately other therapeutics worked as well, the vaccines were developed quickly and had very high efficacy, and subsequent strains of Covid mostly became more mild (this wasn't a guarantee, regardless of how insistent people may be).
The drug works, it reduces hospitalization and mortality in people who take it, but we don't actually need it for the most part so these studies are just further confirmation of that.
And while I understand the "see what sticks" mentality, drastically increasing the mutation rate of the virus responsible for a global pandemic... seems to me like it's on the order of "maybe we start dropping nukes, that'll slow the spread".
It doesn't in vaccinated people - but it did in unvaccinated people. There was no guarantee that we'd develop a vaccine in any reasonable amount of time, and reducing deaths and the overburdening of our limited health care resources was the #1 goal in 2020.
It's a bit ambiguous to me in that Merck originally argued that molnupiravir would be less susceptible to resistance for Reason A, then now that the bad news is coming out, we have a plausible sounding explanation Reason B that it does have a weakness to resistance. The fact that the stated reasons can just flip on a dime like that bothers me intellectually.
But on another level the whole issue is less relevant than the fact that the West has already been recommending Paxlovid over molnupiravir anyways, not because of resistance concerns but because Paxlovid works better and has less side effects, etc.
Here's a boring 1h video discussing all kinds of hand poses. (I think the channel got a bunch of backlash for pro-colonialism or something like that.) It appears I watched half an hour of it, maybe as a form of induced delirium.
These old units are common in engineering applications across North America, the US especially. When I was in chemical engineering school, we had to be effectively bilingual in terms of units. We even used some bizarre units like the lb-mol, defined as the number of atoms in 12 lbs of carbon-12 (ie 454 times more than a gram mol).
My understanding is that the standard model likely does predict static electricity, but since it's a phenomenon bigger than a few molecules, we have no way to actually run the simulation. The physics is willing but the computers are weak.
Fluorine forms the strongest bonds to carbon that are available (much stronger than a carbon-carbon or carbon-hydrogen bond, also stronger than carbon-chlorine). It acts like an immovable stub preventing further reactions, which is great for materials like nonstick coatings, but also prevents natural breakdown in the environment.
A simulator like this would require a lot of simplifications: a table of template chemical reactions and thermochemical mixing rules which are bound to be wrong in edge cases. It doesn't take many reagents before you hit a combinatorial wall where the combination has never been tried in practice, and while it could be estimated with quantum mechanics in theory, that gets expensive.
This is what popped into my head as well. In situations where you have multiple phases so that Phase A is required to reach a certain temp before adding Phase B, but Phase A must reach a specific temp then cool back down within a range before adding Phase B. Once Phase A and Phase B are combined, do you stir, blend, whisk, whip, etc? What happens when you need to use 0.14 grams, but add 0.15 grams?
Each little variance as minor as they may seem will affect the results. Making something one time is a definite achievement. Making the same thing multiple times, or successfully scaling the formula up/down is also noteworthy as well. (assuming doing things by hand as amateur from the title implies to me)
All of that to say, it would be MFing impressive if some sort of simulation could accurately demonstrate these kinds of variations in the results. Could be useful to try to recreate where something went wrong as a sort of reversing/debugging the result.
sorry i'm not sure whos comment to reply under - so i'll just stick it here - thanks for the explanations :) I am ignorant of maths, chem, comp science...so my question was posed purely for my own knowledge.
What i described could be likened to raytracing for light...but for chemicals...
If something like this is possible, would there be no use or other limitations to making a "descriptive" simulation. ie. libraries of animated reactions? Or is this what comicjk is refering to?
assuming the kinds of experiments listed in this article will turn out similar results if performed with similar materials?
It's more like, we can do raytracing, but what's really needed here is simulating actual photons scattering, complete with all the quantum effects that involves.
That's a good response. I was trying to figure out how to reply, but didn't like anything I came up with. I would honestly have no idea on how to program a computer to simulate the individual atoms and the activity of their electrons. How to tell the computer when two atoms can bond with the sharing of an electron, how to tell it when a covalent bond can be made, etc. That's probably a lot to do with I barely understand it myself.
Also, I'd assume the the sim would assume all equipment is 100% pollutant free, all ingredients are 100% pure so that there is no adulterants being introduced to the formula. Also a room of perfect humidity and temperature etc.
Only if you have something nearby that needs low-grade heat, like warming buildings. Waste heat is a diffuse source of energy that's not worth the infrastructure cost of transporting more than a few miles.
Too expensive, too easily degraded by minor impurities in the fuel, not improving nearly as fast as batteries (their main competition). Using rare materials more efficiently would definitely help with the cost problem.
I'd guess the one important application of fuel cell tech that people often appear to forget is going to be long haul trucks where it'll replace the diesel power train.
It's a big chunk of overall land transport that IMO in the long-term won't have other technologically/economically viable options besides the fuel cell.
Rail doesn't serve the last few miles to the destination.
Electric trucks are viable for short distances.
Trucking dozens of tons of cargo over distances > 500 miles isn't going to roll well with carrying another 3-5 tons of battery. And having to recharge that at 2 MW every now and then would require a very reliable/available and ubiquitous high power charging infrastructure.
I think about 2 ton battery is closer to truth for a truck that has about 400-500 mile range. Less, if you consider all the heavy diesel engine and transmission parts an electric truck is not going to need.
Charging at the starting point while loading, at (mandatory) breaks and at the destination should be enough; a BEV truck done right shouldn't require extra waiting time.
> I think about 2 ton battery is closer to truth for a truck that has about 400-500 mile range.
A Tesla Model 3 has a ~ 500 kg battery and weighs ~ 2 tons. With the drag coefficient of a truck being significantly greater and its gross weight amounting to ~ 15-20 times that of the Model 3, I'd say your 2 tons are way off. Sure, the truck will go slower than the Model 3, but still.
Also electric motors do weigh a few kg's as well, so I'd guess the less heavy drive train of the electric vehicle isn't going to save all that much weight.
> Charging at the starting point while loading, at (mandatory) breaks and at the destination should be enough; a BEV truck done right shouldn't require extra waiting time.
I think this needs to be compared to the procedure with a diesel truck. The diesel truck needs a few minutes at the gas station to refill and get some Ad Blue or what. Then it's all flexible to go anywhere for a few hundred miles.
Compared to that, your BEV truck is going to need careful planning ahead of charging and any small deviation from the plan is going to be a lot more of a hassle than for the diesel truck.
Point being. Nope, there aren't many H2 gas stations either as of now. But once there are, the refuelling of a FCEV will be much more similar to that of a diesel/gasoline vehicle than that of a BEV, aka more convenient/resilient.
"A Tesla Model 3 has a ~ 500 kg battery and weighs ~ 2 tons. With the drag coefficient of a truck being significantly greater and its gross weight amounting to ~ 15-20 times that of the Model 3, I'd say your 2 tons are way off. Sure, the truck will go slower than the Model 3, but still."
A truck should consume about 5x than what a Model 3 LR does. 5x 445 kg (actual M3LR battery weight) is 2225 kg.
A shipping container measures w x h: 2.438 x 2.591 m
The total height of the truck will be > 3 m so we're talking about 2.438 m x 3 m projected area. That's ~ 7.3 m2
A Tesla S apparently has 0.562 m2 drag area so let's assume 0.6 m2 for the Model 3. [1]
This amounts to a factor of Semi to Model 3 of:
7.3 m2 * 0.36 / (0.6 m2 * 0.23) = 19
So at the same speed the aerodynamic drag of a truck will be almost 20 times that of a Model 3. Yes, a truck typically drives slower and speed goes into calculation of engine power at a power of 3. But the truck would have to go slower than the Model 3 by a factor of 19^(1/3) ~ 2.7 to have roughly the same drag.
If the Model 3 drives at 150 kph and the Semi at 100 kph, the Semi still has more than 5 times the aerodynamic drag.
And you'll have to add friction to that and losses for accelerating the greater mass. (I doubt regenerative braking will scale well with increased vehicle mass)
Your drag area figure for Tesla Semi is absurdly high.
Note that drag area is cross-sectional area times drag coefficient. If your numbers are otherwise correct, Semi's "drag area" should be 0.36 * 7.3 m^2 = 2.62800 m^2.
2.62800 m^2 (Semi) / 0.562 m^2 (Model S) is approximately 4.68. So I think 5x energy consumption is completely feasible.
> I doubt regenerative braking will scale well with increased vehicle mass
Why would that be an issue? 500 kWh magnitude battery can absorb about 7x power compared to a Model 3 LR AWD battery. Regenerative braking is probably only ever issue when the battery is somewhere above 95% full.
> Your drag area figure for Tesla Semi is absurdly high.
Don't think so. But I made a different error. See further down.
Drag equation [1]:
> FD = 1/2 * rho * u² * cD * A
> The reference area A is typically defined as the area of the orthographic projection of the object on a plane perpendicular to the direction of motion.
So A in the case of a truck carrying a standard container cannot be smaller than the section of the container. And because the container cannot hover millimetres above the ground but must rather be carried at a height of at least half a metre you'll have A > greater than the cross section of the container.
Which is what I calculated above.
I did make an error though by multiplying the drag area of the Model S by the drag coefficient, since the 0.562 m² already takes the coefficient into account.
So you're right, the factor Semi/Model S is ~ 4.68 based on the numbers I assumed.
It does look more feasible indeed based on this number.
Yet I'm still sceptic a battery 5 times larger will suffice because of higher friction and because I doubt regenerative braking will recover the same proportional amount of energy for the Semi as for the Model S.
Let's see. Decelerating the 20,000 kg Semi going at 100 kph at mild 0.10 g requires a force of 0.1 * 9.81 m/s² * 20,000 kg = 19,620 N.
At a velocity of 100 kph that equals (not taking drag and other friction into account) an initial (lossless) braking power of 545 kW that could be regained by regenerative braking. Okay, could be feasible as well, if charging can be ramped up to this rate within the fraction of a second.
If you brake at 0.5 g though, you'd have to suddenly feed in the ball park of 2 MW into the battery. Not sure that's possible.
> If you brake at 0.5 g though, you'd have to suddenly feed in the ball park of 2 MW into the battery. Not sure that's possible.
MCS [0] charging standard goes up to 3.75 MW. I don't think Semi can charge at that power, but 2 MW — why not.
Of course, the catch is that the higher the battery state of charge (SoC), the lower the charging current can be. 2 MW might not be possible, say, somewhere above 50-80% SoC.
That ignores all the existing infra (or lack thereof).
Fossil fuels are very energy dense, and we still have tons of truck stops everywhere - and need them!
Last mile, most dropoffs are not going to have power infra to allow MW+ Charging of every truck that shows up, at least not without a lot of time to upgrade. And many won’t want to even try, as they’re paying the logistics companies so they don’t need to deal with stuff like that.
Even distribution centers would struggle (capex wise), as we’d be talking 100s of megawatts at least of extra load, possibly giggawatts.
There has been no more meaningful progressive in batteries in over a decade. The energy density of batteries today (~265 Wh/kg) is marginally better than where it was in 2012 (~250 Wh/kg). It's been entirely a function of cost reduction. If this continues, people will need to stop talking about "rapid advances" in batteries and instead talk about stagnation.
The report is wrong. It doesn't even make sense since there is clearly a dot above 200 Wh/kg in 2012. Meaning the graph is only claiming a 40-50% improvement in the last decade.
But regardless, the report is wrong because we most definitely had reached 250 Wh/kg by 2010. Panasonic mass produced a cell with those specs start in 2009: https://news.panasonic.com/global/press/en091218-2
Furthermore, there is no way of buying that 300 Wh/kg cell shown on the chart. No seems to have ever found one available as a commercial product. Meaning it is likely an experimental cell that never made it to production.
Another failed miner who is rather well known: Mark Twain. His comic memoir about trying to get rich as a young man in the Nevada and California mines is called "Roughing It."
