Thanks for these insights! Do you have any material or sources for a layman to learn more about this and where these numbers come from?

Aren&#x27;t the elephants and whales orders of magnitude better than us at that though (they have roughly as many cancer as we do, but with respectively x100 and x1000 times as many cells.Or is it the second layer that works better for them?

Our L1 defense is actually incredibly good. A human will undergo about 10^16 cell divisions over the course of their lifetime. Around 10^3 to 10^6 of those divisions will result in a mutation that gets past the L1 defenses and need to be eliminated by the T cells. It&#x27;s not generally easy to make dramatic improvements to something with a 99.9999999% success rate.The immune system is pretty good too, which means any given improvement to the replication system is, all else being equal, probably going to prevent mutations the T cells would already handle. If you need to do the research to figure out what&#x27;s getting past the immune system anyways, and improving the immune system is lower hanging fruit, it&#x27;s the logical place to start.

If the organism is too efficient at preventing mutations - it would evolve slower, right?

This is a fascinating niche of evolutionary biology that I have worked in for a while. The short answer is that yes, as far as we can tell all organisms evolve increasingly more efficient replication machinery, however at some point the strength of selection is no longer powerful enough to overcome the strength of genetic drift and some degree of error rate persists. As far as we can tell it seems like population size governs where this balance ends up such that small populations have high mutation rates and large populations have reduced mutation rates. Michael Lynch coined the term drift barrier hypothesis to describe this phenomenon. <a href="https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;23077252&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pubmed.ncbi.nlm.nih.gov&#x2F;23077252&#x2F;</a>

&gt; To me it sounds a tough problem given the permutation and combination of mutation— roughly few trillions.You are right. There is a very good explanation in this comic <a href="https:&#x2F;&#x2F;phdcomics.com&#x2F;comics.php?f=1162" rel="nofollow">https:&#x2F;&#x2F;phdcomics.com&#x2F;comics.php?f=1162</a>

While cancer is caused by mutations in the genome, these mutations in turn produce the unifying property of cancer: unchecked cell replication.Most cell types have systems to safely manage replication. Broadly, there are gas pedals (oncogenes) and brakes (tumor suppressors). A classic oncogene is something like RAS, which activates a signaling cascacde and stimulates progression through the cell cycle. A canonical tumor suppressor is something like TP53, the most frequently mutated gene in cancer, which senses various cellular stresses and induces apoptosis or senescence.Most cancer genomes are more complicated than individual point mutations (SNPs), insertions, or deletions. There are copy number alterations, where you have &gt; or &lt; 2 copies of a genomic region or chromosome, large scale genomic rearrangements, metabolism changes, and extrachromosomal DNA. There is a series on the hallmarks of cancer which is a useful overview [1].All of the mechanisms that intrinsically regulate cell growth would fall under your &quot;L1 defense&quot;. Unfortunately, the idea of reversing somatic point mutations is likely to be a challenging approach to treating cancer given the current state of technology.First, for the reasons above, cancer is often multifactorial and it would be difficult to identify a single driver that would effectively cure the disease if corrected. Second, we don&#x27;t have currently delivery or in vivo base editing technology that is sensitive or specific enough to cure cancer by this means. There are gene therapies like zolgensma[2] which act to introduce a working episomal (not replacing the damaged version in the genome) copy of the gene responsible for SMA. There are also in vivo cell therapies like CAR T which attempt to introduce a transgene that encodes for an anti-cancer effector on T cells. These sorts of approaches may give some insight into the current state of art in this field.Edit: also I should note that the genes involved in DNA repair (PARP, BRACA1&#x2F;2, MSH2, MLH1, etc) are frequently mutated in cancers and therapeutically relevant. There are drugs that target them, sometimes rather successfully (e.g. PARP inhibitors). But the mechanisms of action for these therapies are more complicated than outright correcting the somatic mutations.1. <a href="https:&#x2F;&#x2F;aacrjournals.org&#x2F;cancerdiscovery&#x2F;article&#x2F;12&#x2F;1&#x2F;31&#x2F;675608&#x2F;Hallmarks-of-Cancer-New-DimensionsHallmarks-of" rel="nofollow">https:&#x2F;&#x2F;aacrjournals.org&#x2F;cancerdiscovery&#x2F;article&#x2F;12&#x2F;1&#x2F;31&#x2F;675...</a>
2. <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Onasemnogene_abeparvovec" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Onasemnogene_abeparvovec</a>

I don&#x27;t know the specifics of efforts to &quot;repair&quot; DNA replication repair (if you get my drift), but I suspect there is some effort in that area.There definitely are efforts to correct enzymes involved in tumor-suppression (p53 is probably the best known tumor suppressor protein). e.g., here&#x27;s a study on a small molecule designed to correct mutated p53 <a href="https:&#x2F;&#x2F;pmc.ncbi.nlm.nih.gov&#x2F;articles&#x2F;PMC8099409&#x2F;" rel="nofollow">https:&#x2F;&#x2F;pmc.ncbi.nlm.nih.gov&#x2F;articles&#x2F;PMC8099409&#x2F;</a>

&gt; fixing the enzyme that fixes the wrong copy paste mechanismThe DNA fidelity issues contribute to only some cancers. Many are caused by mutations due to environmental damage and some are caused by viruses. The point is, there&#x27;s a huge variety of reasons for developing cancer. So you cover more cases by developing treatments that are more &quot;universal&quot;.

Would sports also increase the rate of division? Bodybuilding for example you intentionally make micro tears in muscles to get them to repair and grow (lay person, apologies for the mistakes)

It starts with mutations (sometimes accelerated by mutagens (smoke, alcohol, etc) or inflammation (viruses, infections, etc) or just chance (things like asbestos up the division rate by constant physical damage and thus up the probability or an error in copying).But there is much more to it. This is a nice paper for an overview: Hallmarks of Cancer (tng) [0]. It (among others) adds the very important and for years underestimated role of the immune system to the original 2000 paper.[0] <a href="https:&#x2F;&#x2F;www.cell.com&#x2F;fulltext&#x2F;S0092-8674(11)00127-9" rel="nofollow">https:&#x2F;&#x2F;www.cell.com&#x2F;fulltext&#x2F;S0092-8674(11)00127-9</a>

Cancer sucks, I wish all the best towards a recovery.You’d also have to ‘fix’ DNA: unless we can re-engineer a bunch of key enzymes and then re-encode the entire genome (or maybe key parts) with forward error correction without breaking everything else, it might work. You might also break evolution to some degree by making random point mutations less likely.But what I learned so far is that as soon as you’d attempt something like this in bacteria, the fitness advantage from an evolutionary standpoint is negligible compared to the efficiency loss introduced by FEC, so your colony would get outcompeted by other bacteria unless there is a niche your resistant bacteria survive in (high radiation environments?). The efficiency loss induced ‘disadvantages’ would probably be less pronounced in mammals though - If (big if) you manage to not also break anything essential in the wonderful yet surprisingly efficient Rube Goldberg machine that is life.

What I meant was there are collection of genes responsible for error correction. If there is a failure in duplication then these genes have not done their job.Thought experiment, again as a layman, was to see if these genes responsible for error correction at the base level can be fixed or bolstered and that will act like a cancer vaccine. But looks like from other comments that this is even more harder!

Cancer sucks and I wish your father the best.Also not a doctor or microbiologist, but just wanted to share my layman’s guess on why fixing enzymes will not completely solve the issue: there’s 2 strands of DNA and to fix the broken (mutated) strand you need to have one correct template strand intact so you know what it should be fixed into. It could be the nucleotides swapped places between strands or are deleted completely or otherwise both mutated, which would mean any repair will not revert the sequence to what it used to be.The other comments so far are probably more informed.

&gt; fixing the enzyme that fixes the wrong copy paste mechanism. Or making the enzyme get more efficient and powerfulIf we had the tools to easily do that we’d practically be gods.

L2 is being prioritized because our L1 defences are already very good. We are a long-lived species, and our natural ability to fix errors is so good that it is hard to improve upon. Maybe the long-lived tortoises or whales can do it better, maybe. But we have &quot;several nines of reliability&quot; there.OTOH our L2 isn&#x27;t that good, mammals in general (with some notable exceptions such as bats, whales and naked mole rats) are prone to cancer in their older age. There probably is a lot of relatively low-hanging fruit there.If you think about it - individual cells aren&#x27;t very precious and if some of them gets FUBARed by something (a virus, radiation or chemical insult), it is better to whack it and reuse the proteins to build a new one, if possible, instead of wasting time and resources on reconstruction of a total wreck.Which also means that some research into replenishment of stem cells is necessary - and this is, IMHO, the really underfunded part of the whole thing. We lose a lot of stem cells as we age. Maybe we don&#x27;t have to.

I know this is a bit off topic, but have you ever thought about why steroids and other forms of doping are not a free lunch? Why can&#x27;t we just inject an external chemical to boost our strength for free without any side effects?If steroids worked, everyone would be constantly injecting them. It would be like drinking coffee.And that is the reason why steroid injections are harmful. If there is a free lunch, the human body will simply produce the optimal amount of steroids on its own until the Pareto frontier is reached and a tradeoff needs to be made.Where does the body get the materials to form the steroids? From your diet. So the primary intervention is always a healthy diet and an active lifestyle. You know, the boring things that parents drill into their children.It&#x27;s valid but &quot;medicine&quot; that has only upsides and no downsides isn&#x27;t medicine, it&#x27;s diet.

&gt; But I was curious if there is working happening on L1 defence — fixing the enzyme that fixes the wrong copy paste mechanism. Or making the enzyme get more efficient and powerful. Is that line of thought even valid?Mutations in general are not the defining quality of cancer. It&#x27;s mutations in these very L1 safeguards. There are several such safeguards and a cell needs several mutations in those to become malignant. Eg. <a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;P53" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;P53</a>Correcting genes only works in certain conditions (e.g. limited single strand breaks), in a narrow time frame during cell division, safeguards rather trigger cell suicide, or if that fails they mark the cell for destruction by immune cells. A cell can&#x27;t fix DNA which made it through cell division once, because it got nothing to proof-read against.After the safeguards are gone, everything goes and genetic diversity increases quickly within each tumor. This diversity is what&#x27;s making cancer treatment hard. At some point there won&#x27;t be a shared vulnerability in all malignant cells. The repair mechanisms are working in favor of the cancer now. For example, with radiation therapy you preferably want to induce DNA double strand breaks, because cancer cells can&#x27;t repair those. Otherwise you need to increase the radical burden enough to overwhelm repair, but migrating radicals may damage distant cells, too.I presume you could hypothetically inject mRNA of a working safeguard gene (eg. P53) into all cells (at some point cancer cells can&#x27;t be selected exclusively, since they lost identifying marks and present as stem cells), so the functional enzyme or transcription factor is forced to be built inside. I am sure people are trying this right now. However, the inner workings of cells on a molecular level are insanely complex and our understanding is only scratching the surface. As with P53, you have a transcription factor, which means it&#x27;s modifying gene expressions elsewhere. It&#x27;s only a small part of a complex regulatory cascade. I doubt there is a safeguard target, which can easily be injected without considering the precise timing and environment within that safeguard cascade in the cell. Of course, the rest of the safeguard system needs to be present in the cell to begin with. Mind you, you don&#x27;t want to cause cell suicide in healthy cells, so you want to restore the function of whole selective complex.Then there is the question of delivery. Can you deliver eg. the mRNA to every cell without raising suspicion of the immune system? With the COVID vaccine, the enabling breakthrough was the delivery vehicle, not as much the mRNA part. Can you even reach the cancer cells at all? Cancer cells are frequently cloaked, shadowed or cut of by senescent, or necrotic cells, or acquired unique ways of metabolic adaptations. A bit similar to bacterial persistence, like M. tuberculosis, which can evade bodily and chemical defenses for decades.The take-away: Life is complex beyond comprehension! Despite simplifications taught in schools and reductionist zeitgeist, we actually know very, very little about what&#x27;s going on in genetics and molecular biology, most medical knowledge is empiric guessing instead of explanatory understanding.

I have been working with my dad on his cancer treatment since last year. My interest in the topic has only peaked ever since.(Disclaimer- I am an engineer and not a microbiologist&#x2F;doctor)Mutations and wrong copying of genome happens all the time in the body and some enzyme has the job of correcting the mutated genes so it doesn’t get into the system. Level 2 defence is T cells killing it as identified as foreign body.Thing that baffles me is that I see most work happening to eliminate tumor. To me it sounds a tough problem given the permutation and combination of mutation— roughly few trillions.But I was curious if there is working happening on L1 defence — fixing the enzyme that fixes the wrong copy paste mechanism. Or making the enzyme get more efficient and powerful. Is that line of thought even valid?

It looks like they screened 45 strains of bacteria to find 9 that passed their safety tests, and then only one of those had a 100-percent response. The sample size is also small: 5&#x2F;5 sounds a lot less impressive than 100%. I&#x27;d expect the true response rate to be substantially lower (&quot;winner&#x27;s curse&quot;).The bar for an acceptable side effect profile in an FDA-approved drug would also be a lot higher than &quot;five genetically near-identical mice did not show evidence of pathology in a single study.&quot;I&#x27;m not saying this work is bad (skimming, it seems fine for what it is, though haven&#x27;t read in detail), but it&#x27;s quite preliminary if we&#x27;re talking about developing a medical treatment that could eventually be deployed in humans. There&#x27;s a reason it ended up in a mid-tier microbiome journal.

What&#x27;s the difference between a good and not-good reason to not go to market?

From what I saw, there are very few experimental therapies that claim 100% efficacy, no side effect, no patent on complex engineering process to produce the drug.I agree with GP that it is very notable.I mean if it works on humans, which is not a stretch, colorectal cancer is done. It&#x27;s huge.

The catch is that there are thousands of promising therapies in animal models&#x2F;pre-human testing. A very very tiny fraction of them will ever make it to market for a variety of both good and not-good reasons.

Certainly good news for mice.I wonder if anyone has tried to engineer a mouse that lives forever by applying all these life enhancing mouse therapies at once.

Very small sample size and only in mice. Most experimental treatments don&#x27;t make it to the clinic, and this could be another in a long list of exciting early results that don&#x27;t lead to anything.
However...
These results seem too good to be true, but the paper actually seems pretty good. There have been other attempts to use anaerobic bacteria to treat cancer (as solid tumors tend to be hypoxic environments), this is the first I&#x27;ve seen that interacts with PD-1&#x2F;PD-L1 pathway. This is the most promising cancer treatment I&#x27;ve seen in quite a while.
Maybe we can be cautiously optimistic!

I hesitate to link directly to a single podcast dealer, but if you search for Dave Ricks, CEO of Eli Lilly&#x27;s podcast with Stripe brothers, there&#x27;s some alpha in there.

A few years ago we also had 100% effective medication with zero side effects. Now birth rates are down, mortality up, and young people misteriosly die from cancer and hearth attacks.

100% response, zero side effects?This sounds like world changing news. Can anyone with domain expertise explain the catch, if any?

I&#x27;m untrained in this area but it sounds like similar ideas I heard of in the last decade where injected tumor with virus (polio?) that quickly kills the tumor but then is easily cured in the rest of the body since we have solutions to them. Is this a similar situation?

Still the idea is beautiful. Since tumors are oxygen-deficient and suppress the immune response, anaerobic bacteria would proliferate there, and wreak havoc, while in the healthy parts of the organism they would be rapidly eliminated. Additionally, since the bacteria accumulate in the tumor, and the immune system has just responded to their invasion, T-cells will flock to the tumor, destroying what remains of it in due course.As they say, &quot;the fame of a mathematician is measured by the number of poor papers&quot;, because pioneering works are often awkward, treading completely unknown ground. Maybe the same applies to biology sometimes?

Figures 2 and 3 seem to be different experiments, with n=3 and n=5, respectively. Both showed 100% survival. Obviously very small sample sizes, but still promising.

Yes in figure 2 it&#x27;s 3 mice, next figure 3 they also have 5 (panel e)

Interesting article, but in the full paper their key figure (Fig 2) shows their treatment group of n=3 mice completely responded to the bacterial treatment, but their methods say they treated n=5 mice? Could be an honest mistake but that’s a little concerning for data manipulation.Also agree that using a PD-L1 mab feels like it’s for show especially considering the cancer model they’re using (Colon-26) was shown to be substantially less responsive to PD-L1 inhibitors…Not the world’s best paper imo

<a href="https:&#x2F;&#x2F;www.tandfonline.com&#x2F;doi&#x2F;full&#x2F;10.1080&#x2F;19490976.2025.2599562" rel="nofollow">https:&#x2F;&#x2F;www.tandfonline.com&#x2F;doi&#x2F;full&#x2F;10.1080&#x2F;19490976.2025.2...</a>
Link to the research article. 
Are you referring to the news release?

I was impressed by the clarity and terseness of this little article.
Is it common practice in the case of scientific articles coming from Japan?

I know we&#x27;re not supposed to edit titles, but I&#x27;m glad the submitter added &quot;in mice&quot;. It avoided me quite the disappointment!

Yes it is amazing! Solid tumors tend to be poorly oxygenated, as they don&#x27;t have a good network of blood vessels to supply them. The bacteria in these experiments can only live in low oxygen environments, so they will multiply in the tumor and die in any other part of the body they end up in. It&#x27;s a clever idea, hopefully it will be successful.

Even lay-er person, but maybe the specificity is not that impressive in mice? Perhaps when you scale to more complex animals it is inevitable to see false positives (detrimental effects to healthy cells)?

So maybe I am too much a layperson here, but even without any direct therapetutic effects, it is pretty remarkable to have an easily scalable mechanism to get self-replicating agents into tumors, but nowhere else, is it not?

It is my wish and my blessing that this is true

&quot;Cancer&quot; is a term that covers a lot of diseases. So there is a lot of research going into a lot of different things, and hence lots of announcements.&quot;Chemotherapy&quot; again is a loaded term covering a lot of different drugs, drug combinations, protocols and so on. So yeah, a lot of cancer treatment us &quot;chemo&quot; - but today&#x27;s chemo is far removed from 2000 chemo.5 year survivability has increased tremendously over the last decades. We&#x27;re not talking 0.5% here, breast cancer for example has gone from 72% to 93%. Early detection of prostrate cancer has near 100% survivability.But you&#x27;re right, improving survivability doesn&#x27;t make for sexy headlines. Yes there&#x27;s a social media appetite for &quot;breakthroughs&quot;, but the underlying &quot;boring&quot; stuff is doing well, and getting better all the time. It&#x27;s just not &quot;news&quot;.

I&#x27;m not a doctor, but in some fairness, I think there has been a lot of progress in chemotherapy and radiation. &quot;Increasing 5-year-survivability by 0.5%&quot; doesn&#x27;t make a fun sexy headline, but that&#x27;s still an achievement that required a lot of hard work and enough of those happening still adds up.I agree with your overall point though; it&#x27;s a little annoying that every few weeks we hear about a new experiment that seems to indicate that we&#x27;ll have a radically new and effective form of treatment for cancer only for it to never materialize.

There are a few factors at play here, I think* Many breakthroughs from the first research stages never make it into medical application.* Many breakthroughs are touted as some kind of &quot;novel treatment&quot;, but when they get into the hands of the doctor, they talk about it as chemotherapy, because it kills cancer cells. So you might not even notice that you&#x27;re getting something novel.* Many breakthroughs take decades until they actually land in mainstream treatment.* Many breakthroughs are specific to some kinds of cancer.That said, in most developed countries, survival rates&#x2F;times for cancer have been steadily improving for decades.It&#x27;s a bit like with solar cell and battery tech breakthroughs: you hear about them all the time, but it takes 20 to 30 years until they make it to production. But both have been improving steadily for an impressively long time.

Yes, you only hear about the initial breakthrough. The regular news doesn&#x27;t report the trials unless things go spectacularly wrong. And you don&#x27;t hear about the successes using the therapy on patients where traditional treatments didn&#x27;t work. And you don&#x27;t hear about the treatments successful enough that they replace the traditional treatments. But they are there, being used and saving lives. A family member&#x27;s prostate cancer metastasized after 20 years of hormone therapy, they refused chemotherapy, and had their life saved by radioligand therapy, a treatment not available just months earlier. No side effects beyond a dry mouth, and now off the more debilitating hormone treatments.

<a href="https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Monoclonal_antibody_therapy" rel="nofollow">https:&#x2F;&#x2F;en.wikipedia.org&#x2F;wiki&#x2F;Monoclonal_antibody_therapy</a>These are not your parents cancer treatments.

There&#x27;s also AIDS that I heard have been practically cured. Since then I have became less pessimistic about drug progress

Keytruda (pembrolizumab) has been pretty monumental.

I have never once seen a promising cancer treatment I&#x27;ve heard of on the news help people. You hear about the breakthrough treatments all the time, but when people get cancer, all you ever hear about is people getting chemotherapy and radiation. Same old scary shit.Well, I guess Leukemia has been somewhat cured I heard, so that&#x27;s pretty huge. When I was a kid it was a death sentence IIRC.

In case people don&#x27;t want to click on a random link, it is actually quite interesting:Crocodile blood antibiotics hopeScientists are catching crocodiles and sampling their blood in the hope of finding powerful new drugs to fight human infections.Even horrific fighting wounds on the animal heal quickly

<a href="http:&#x2F;&#x2F;news.bbc.co.uk&#x2F;1&#x2F;hi&#x2F;health&#x2F;4155522.stm" rel="nofollow">http:&#x2F;&#x2F;news.bbc.co.uk&#x2F;1&#x2F;hi&#x2F;health&#x2F;4155522.stm</a>

Keytruda, pembrolizumab, (what he probably received) can only do so much. If it was in his GI tract it was also elsewhere in multiple places. The PD-L1 drugs at this point have more than 400k patients treated, with decent efficacy. I&#x27;m sorry for your loss. If his melanoma had metastasized to his GI tract it was too late for anything except palliative care.This drug has been used in a huge number of patients for more than 11 years; the next gen of drugs is currently being used. I&#x27;m sorry for my curt style of writing, but - people like your father have helped pave the way for that next generation of drugs by constraining clinical trial designs.

In my dad’s case- he had gastric melonama. We surgically removed it and as consolidation We administered pd-L1 Immune checkpoint inhibitor. Melonama recurred again in 6 months time. This time in esophagus.As an engineer I think all drugs tested and efficacies studied are on statistically not so significant data points. Given the permutations and combinations far exceed the clinical trials available and hence everything post clinical trial is also just an extended trial.Wonder How to fix this? I am assuming heLa cells etc are also not the right test setup to have better test results.

&gt; Sorry, as someone in this field, this is bullshit. It is in mice.Nice to hear an expert opinion. Let&#x27;s hope your comment goes back to black. I have a lot of question!&gt; This is a nothing burger.Is it enough for a bread-mayo-bread sandwich? Lettuce?IIUC the bacteria makes the cancer disappear for two weeks, until they end the study and kill the mice. (IIUC this is timeline is usual for very early studies.) They tried other bacterias and one of them made the cancer disappear for a few days, so I&#x27;m worried about the long time efficiency of this method.Is injecting the bacterias a second time as efficient as the first time, or the inmune system kills the bacteria before they hurt the cancer?What happen in case of metastasis? Each one must be injected with the bacterias or they will jump and make all of them disappear?Does the bacteria infect other organs and kill you? Is there a good antibiotic in case the bacteria cause problems?They used cancers that were 200mm3 (i.e. like a sphere of 7mm = 1&#x2F;4 inch). What happens in bigger cancers? Does bigger cancer have better irrigation and make it more difficult for the bacteria to survive? What happens to tiny hidden metastasis (that probably still have good enough irrigation)?

Seems like a very interesting approach, even if it’s early stage.&gt; Many tumor types are not PD-l1 positive.
&gt; Doxy is an ancient SOC chemo. This is a nothing burger.Meh the research didn’t say those were state of the art, but that they were “common” treatments. In other words a baseline for a presumably cheap and well studied animal surrogate.&gt; CAR-T is ahead of the game, and will be the ultimate winner here as it grows to scale.Last I read up on it last year CAR-T treatments struggled with solid mass tumors.Many cancers don’t have unique proteins for CAR-T to target (similar to the pd-l1 issue).Then CAR-T struggles getting the modified T cells into the solid mass tumors en masse. Interestingly this approach actually makes use of the tumor environment rather than be hindered by it.

Sorry, as someone in this field, this is bullshit. It is in mice.Several things trigger my bullshit meter. Quote:&quot;This dramatically surpasses the therapeutic efficacy of current standard treatments, including immune checkpoint inhibitors (anti-PD-L1 antibody) and liposomal doxorubicin (chemotherapy agents)&quot;PD-L1 monoclonal antibodies are only effective against cancers that are, you guessed it, PD-L1 positive. At high percentages, ranging from 1 to 50%. Are these authors even familiar with the state of the art when it comes to cancer medications? Mouse tumors do not equate to people tumors. Many tumor types are not PD-l1 positive.Doxy is an ancient SOC chemo.This is a nothing burger.Give me phase II&#x2F;III clinical trials, and then let me know what their PFS&#x2F;OS was after 5 years. and what the medians were at 3- and 5-years. Also, ORR and CR and needed.CAR-T is ahead of the game, and will be the ultimate winner here as it grows to scale.

Previously: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46283089">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=46283089</a>

Who knows how much knowledge we eradicated due to not bothering with climate change and just letting species go extinct.
Thankfully these were still here for this discovery.

This segment about the mechanism is simple and very profound. I wonder if any cancer researchers here could comment on its universality across various types of cancers:&quot;Tumor-Specific Accumulation MechanismE. americana selectively accumulates in tumor tissues with zero colonization in normal organs. This remarkable tumor specificity arises from multiple synergistic mechanisms:Hypoxic Environment: The characteristic hypoxia of tumor tissues promotes anaerobic bacterial proliferationImmunosuppressive Environment: CD47 protein expressed by cancer cells creates local immunosuppression, forming a permissive niche for bacterial survivalAbnormal Vascular Structure: Tumor vessels are leaky, facilitating bacterial extravasationMetabolic Abnormalities: Tumor-specific metabolites support selective bacterial growthExcellent Safety ProfileComprehensive safety evaluation revealed that E. americana demonstrates:Rapid blood clearance (half-life ~1.2 hours, completely undetectable at 24 hours)Zero bacterial colonization in normal organs including liver, spleen, lung, kidney, and heartOnly transient mild inflammatory responses, normalizing within 72 hoursNo chronic toxicity during 60-day extended observation&quot;

As a rule of thumb, it’s best to assume that all studies like this are in mice or rats unless the headline specifically says “in human trials”.Murine studies are a dime a dozen and therefore it’s the default assumption when reading research papers. When human trials commence the fact that it’s in humans is a big part of the research and therefore paper titles.

Or alternatively, some marker to indicate the presence of an “only in mice” comment

Maybe we find out why things work in mice and not us.

I would be in favor of adding a standardized [in mice] to the titles of all HN submissions about medical breakthroughs. Most of them end up being in mice and many do not reproduce in humans. It would help, at a glance, to know how significant a study&#x27;s results are.

Of course it was done in mice, tests with animals are obviously mandatory before human trials.

They&#x27;ve given more lives to humanity than humanity itself (j&#x2F;k)

No, it is not relevant in the least. Murine studies are a standard practice on the path to human trials.Your link is not even about animal studies. It is about a petri dish.

Ah, I was wondering if it was one of those. And of course there&#x27;s a relevant xkcd [0][0] <a href="https:&#x2F;&#x2F;xkcd.com&#x2F;1217&#x2F;" rel="nofollow">https:&#x2F;&#x2F;xkcd.com&#x2F;1217&#x2F;</a>

Just FYI, this study was done in mice.

I have significant experience in polymer chemistry, as an experiment, I decided to ask gemini some very specific questions to try and back it into a corner, so to speak. It blew me away with the answer, discussing quite a bit of info I was not even aware of.

&gt; I&#x27;m not against AI summaries being on HN, however, users should verify and cite sources so others can verify.I don&#x27;t see how they contribute anything to a discussion. Even a speculative comment organically produced is more worthwhile than feeding a slop machine back into itself. I don&#x27;t go out for coffee to discuss LLM summaries with friends, and I can&#x27;t imagine why anyone would want to do that here.Earlier today I asked Gemini Pro to find information on a person&#x27;s death that was turning up nothing for me otherwise, and it just imagined finding verbatim Obituary quotes in every source, cobbled together vaguely related names, plausible bits and pieces from wherever, almost like it was 2023 again.It ain&#x27;t search, and it ain&#x27;t worthwhile; I&#x27;d much rather someone ask an llm the question and then post a question out of curiosity based on it, but without the summary itself

I&#x27;ve had quite good luck asking Gemini and ChatGPT to include links to research papers for every claim they make. Not only can I review at least the abstracts but I find when I do this, they&#x27;ll retract some of the hallucinations they&#x27;ve have made in prior messages. It almost seems (and maybe they do) in their web searching tools, reread the content they include. Thus, greatly reducing errors, with minimal extra effort on my part.

I know hearing this gets old, however, please review sources outside of LLMs for accuracy. LLMs take a whole bunch off stuff from all over the internet and distill it down to something you can consume. Those sources include everything from reddit to a certain de-wormer that folks still think treats COVID (side note: I&#x27;ve a few long COVID victims in a support group I am in, and they are not happy about the disinfo that was spread, at any rate)...LLMs&#x2F;&quot;AI&quot; does not and cannot innovate, it can only take all existing information it knows, mash it all together, and present you with a result according to what the model is trained on.I&#x27;m not against AI summaries being on HN, however, users should verify and cite sources so others can verify.However, I&#x27;m just a normal nerd that wants to fact check stuff. Perhaps I&#x27;m wrong in wanting to do this. We&#x27;ll see.

&gt; Could it be that this organism switches to anaerobic respiration when it finds itself inside cancer tissueUnlikely. The leading hypothesis is that mitochondria are a part of the apoptosis cycle, so cells need to disable them to become cancerous. This is called the Warburg effect.There are several drugs that target this mechanism, inhibiting the anaerobic metabolism. They are effective initially, but cancers always find ways to work around them.

I&#x27;m happy to wait for the experimentation to weigh in for this one dawg

&quot;according to Gemini&quot; ........

I believe parent-poster is making a joke at the expense of conspiracy-theorists who believe in lizard-people, rather than endorsing the idea. To paraphrase, something like:&quot;This news article links human survival to something taken from amphibians and reptiles. There are conspiracy-theorists who posit the existence of Lizard people. It would be amusing if those humans discovered this news, and claimed it was a plot by Lizard People to make us dependent on them.&quot;So there, I think I explained the joke... which isn&#x27;t necessarily a good thing. In the words of E.B. White:&gt; Explaining a joke is like dissecting a frog. You understand it better but the frog dies in the process.Now, I could work that new amphibian-connection into another joke... But let&#x27;s face it, it would be &quot;too meta&quot; at this point.

Now I can&#x27;t wait for the conspiracy theory types to say this proves reptilian people theories. Lizard people just giving to help us accept them type of stuff, and maybe prove how the ape descendants need the lizard people.

Gut bacteria from amphibians and reptiles achieve tumor elimination in mice

Beginning January 2026, all ACM publications will be made open access

Classical statues were not painted horribly

Your job is to deliver code you have proven to work

Virtualizing Nvidia HGX B200 GPUs with Open Source

Using TypeScript to Obtain One of the Rarest License Plates

Valve Is Running Apple's Playbook in Reverse

Hi HN, we’re Sid and Ritvik, co-founders of Pulse. Pulse is a document extraction system to create LLM-ready text. We built Pulse as we realized that although modern vision language models are very good at producing plausible text, that makes them risky for OCR and data ingestion at scale.When we started working on document extraction, we assumed the same thing many teams do today: foundation models were improving quickly, multi modal systems appeared to read documents well, and for small or clean inputs that assumption often held. The limitations showed up once we began processing real documents in volume. Long PDFs, dense tables, mixed layouts, low-fidelity scans, and financial or operational data exposed errors that were subtle, hard to detect, and expensive to correct. Outputs often looked reasonable while containing small but meaningful mistakes, especially in tables and numeric fields.A lot of our work since then has been applied research. We run controlled evaluations on complex documents, fine tune vision models, and build labeled datasets where ground truth actually matters. There have been many nights where our team stayed up hand annotating pages, drawing bounding boxes around tables, labeling charts point by point, or debating whether a number was unreadable or simply poorly scanned. That process shaped our intuition far more than benchmarks alone.One thing became clear quickly. The core challenge was not extraction itself, but confidence. Vision language models embed document images into high-dimensional representations optimized for semantic understanding rather than precise transcription. That process is inherently lossy. When uncertainty appears, models tend to resolve it using learned priors instead of surfacing ambiguity. This behavior can be helpful in consumer settings. In production pipelines, it creates verification problems that do not scale well.Pulse grew out of trying to address this gap through system design rather than prompting alone. Instead of treating document understanding as a single generative step, the system separates layout analysis from language modeling. Documents are normalized into structured representations that preserve hierarchy and tables before schema mapping occurs. Extraction is constrained by schemas defined ahead of time, and extracted values are tied back to source locations so uncertainty can be inspected rather than guessed away. In practice, this results in a hybrid approach that combines traditional computer vision techniques, layout models, and vision language models, because no single approach handled these cases reliably on its own.We are intentionally sharing a few documents that reflect the types of inputs that motivated this work. These are representative of cases where we saw generic OCR or VLM-based pipelines struggle.Here is a financial 10K:
<a href="https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example1">https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example1</a>Here is a newspaper:
<a href="https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example2">https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example2</a>Here is a rent roll: <a href="https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example3">https:&#x2F;&#x2F;platform.runpulse.com&#x2F;dashboard&#x2F;examples&#x2F;example3</a>Pulse is not perfect, particularly on highly degraded scans or uncommon handwriting, and there is still room for improvement. The goal is not to eliminate errors entirely, but to make them visible, auditable, and easier to reason about.Pulse is available via usage-based access to the API and platform You can try it here and access the API docs here.Demo link here: <a href="https:&#x2F;&#x2F;video.runpulse.com&#x2F;video&#x2F;pulse-platform-walkthrough-69f9">https:&#x2F;&#x2F;video.runpulse.com&#x2F;video&#x2F;pulse-platform-walkthrough-...</a>We’re interested in hearing how others here evaluate correctness for document extraction, which failure modes you have seen in practice, and what signals you rely on to decide whether an output can be trusted. We will be around to answer questions and are happy to run additional documents if people want to share examples.

Launch HN: Pulse (YC S24) – Production-grade unstructured document extraction

Are Apple gift cards safe to redeem?

Please Just Try Htmx

Jonathan Blow has spent the past decade designing 1,400 puzzles for you

Show HN: A local-first memory store for LLM agents (SQLite)

RCE via ND6 Router Advertisements in FreeBSD

Creating apps like Signal could be 'hostile activity' claims UK watchdog

Slowness is a virtue

Docs: <a href="https:&#x2F;&#x2F;ai.google.dev&#x2F;gemini-api&#x2F;docs&#x2F;gemini-3" rel="nofollow">https:&#x2F;&#x2F;ai.google.dev&#x2F;gemini-api&#x2F;docs&#x2F;gemini-3</a>Developer Blog: <a href="https:&#x2F;&#x2F;blog.google&#x2F;technology&#x2F;developers&#x2F;build-with-gemini-3-flash&#x2F;" rel="nofollow">https:&#x2F;&#x2F;blog.google&#x2F;technology&#x2F;developers&#x2F;build-with-gemini-...</a>Model Card [pdf]: <a href="https:&#x2F;&#x2F;deepmind.google&#x2F;models&#x2F;model-cards&#x2F;gemini-3-flash&#x2F;" rel="nofollow">https:&#x2F;&#x2F;deepmind.google&#x2F;models&#x2F;model-cards&#x2F;gemini-3-flash&#x2F;</a>Gemini 3 Flash in Search AI mode: <a href="https:&#x2F;&#x2F;blog.google&#x2F;products&#x2F;search&#x2F;google-ai-mode-update-gemini-3-flash&#x2F;" rel="nofollow">https:&#x2F;&#x2F;blog.google&#x2F;products&#x2F;search&#x2F;google-ai-mode-update-ge...</a>Deepmind Page: <a href="https:&#x2F;&#x2F;deepmind.google&#x2F;models&#x2F;gemini&#x2F;flash&#x2F;" rel="nofollow">https:&#x2F;&#x2F;deepmind.google&#x2F;models&#x2F;gemini&#x2F;flash&#x2F;</a>