Can someone spell out how this is possible? Do pdfs store a complete document version history? Do they store diffs in the metadata? Does this happen each time the document is edited?
You can replace objects in PDF documents. A PDF is mostly just a bunch of objects of different types so the readers know what to do with them. Each object has a numbered ID. I recommend mutool for decompressing the PDF so you can read it in a text editor:
mutool clean -d in.pdf out.pdf
If you look below you can see a Pages list (1 0 obj) that references (2 0 R) a Page (2 0 obj).
Rather than editing the PDFs in place, it's possible to update these objects to overwrite them by appending a new "generation" of an object. Notice the 0 has been incremented to a 1 here. This allows leaving the original PDF intact while making edits.
1 1 obj
<<
/Type /Pages
/Count 2
/Kids [ 2 0 R 200 0 R ]
>>
endobj
You can have anything inside a PDF that you want really and it could be orphaned so a PDF reader never picks up on it. There's nothing to say an object needs to be referenced (oh, there's a "trailer" at the end of the PDF that says where the Root node is, so they know where to start).
Thanks for the technical explanation! This is pretty fascinating.
So it works kind of like a soft delete — dereference instead of scrubbing the bits.
Is this behavior generally explicitly defined in PDF editors (i.e. an intended feature)? Is it defined in some standard or set of best practices? Or is it a hack (or half baked feature) someone implemented years ago that has just kind of stuck around and propagated?
The intention is to make editing easy and quick on slow and memory deficient computers. This is how for example editing a pdf with form field values can be so fast. It’s just appending new values for those nodes. If you need to omit edits you’d have to regenerate a fresh pdf from the root.
But yeah. It's all just objects pointing at each other. It's mostly tree structured, but not entirely. You have a Catalog of Pages that have Resources, like Fonts (that are likely to be shared by multiple pages hence, not a tree). Each Page has Contents that are a stream of drawing instructions.
This gives you a sense of what it all looks like. The contents of a page is a stack based vector drawing system. Squint a little (or stick it through an LLM) and you'll see Tf switches to Font F4 from the resources at size 14.66, Tj is placing a char at a position etc.
At the bottom of the page there's a link to the pdfresurrect package, whose description says
"The PDF format allows for previous changes to be retained in a revised version of the document, thereby keeping a running history of revisions to the document.
This tool extracts all previous revisions while also producing a summary of changes between revisions."
PDFs are just a table of objects and tree of references to those objects; probably, prior versions of the document were expressed in objects with no references or something like that.
Adding to this, the molecular mechanisms of psychiatric disorders such as depression, bipolar disorder, and schizophrenia are not well understood. People prefer their brains and skulls intact so taking brain samples for study is not possible.
Diagnosis and grading is fuzzy - a cluster behavioral symptoms observed and reported by patients, family members, and clinicians.
Based on some of the genetic evidence, which can also be fraught, it is likely that something like depression is in fact a grouping of a bunch of separate underlying disorders / causes that all result in the same (easily) observable symptoms. Hence why some treatments work on a subset of patients, but are ineffective in others.
I has a Nokia N95. The phone itself was great. The problem was the dearth of apps. Nobody developed anything for windows mobile OS. Maybe Ballmer was not so crazy when he was running around on stage screaming “Developers, developers, developers”.
That is because there has been an absolute massacre in biotech in the bay area. Between tariffs (higher COGS), chaos at the FDA, cuts to NIH funding for basic and translational research, and competition from China biotech ventures are getting squeezed from both ends.
My pet theory is that we are experiencing stagflation, but only people >70 years old have ever really experienced it before, so most people are just scratching their heads wondering how it’s possible that stocks keep going up (inflation) while jobs are disappearing (stagnation). I am most definitely not an economist, nor am I qualified to play one on tv.
> My pet theory is that we are experiencing stagflation, but only people >70 years old have ever really experienced it before, so most people are just scratching their heads wondering how it’s possible that stocks keep going up (inflation) while jobs are disappearing (stagnation).
We do not seem to be technically experiencing stagflation,ir really either half of it, on a national scale, as we appear to still be in a weak aggregate economic expansion and inflation, while higher than the 2% target, is fairly mild at around a 3% annualized rate [0], and, in any case, stocks going up is not inflation (unqualified inflation, which is the inflation part of stagflation, in consumer price inflation, not asset value inflation.)
OTOH, we are in a very weak economy especially outside of the leading AI firms, and there are quite likely both wide regions and wide sectors of the economy which, considered alone, would be in recession, and while inflation is fairly mild, it is high for the last couple decades and being in near-recession conditions. So, for a lot of people, the experience is a something like stagflation (and there are lots of signs that the economic slowdown will continue alongside rising inflation.)
[0] though as economic statistics are only available after the fact, either of these could have changed, but the real defining period for “stagflation” in the US is the 1973-1975 recession, years which saw a minimum of 6.2% inflation (the term was actually coined in the UK for conditions which saw a massive drop in GDP growth rate, fron 5.7% annually to 2.1% in successive years, but not an actual recession, alongside 4.8% inflation.)
>we are in a very weak economy especially outside of the leading AI firms
Isn't that part of the cause? It sucks up so much investment, there's nothing left for anything else. Or at least nothing without such perceived upside.
Either they pull it off and you're replaced by AGI, or they fail to pull it off and you lose your job to the resulting economic implosion.
> It sucks up so much investment, there's nothing left for anything else.
Tariff-inflated input costs combined with weak consumer demand are the reason the rest of the economy is slow, and the reason there aren’t places woth strong and near-term upsides for investment dollars to go. AI being the only thing attracting investment is the effect, not the cause.
My sense is that AI is the one area where boards cannot justify cutting back on investment. If there were no AI boom the rest of the economy would still be getting hammered.
There is still a lot of tech investment, deal making, and hiring going on. It has just left the USA.
The definition of inflation since the 1970s has changed significantly e.g. owner equivalent rent. What many people describe as inflation is the change in base cost of capital purchases which has demonstrably risen at a substantially faster pace than baseline inflation see housing/car/education/healthcare prices.
Economic thought today is that rising asset prices relative to wages are not a sign of inflation. They can be attributed to lower cost of capital, increased dollar production of assets, lower risk profiles, and other aspects. However when observing housing prices and education prices which have seen declining utility over the years - it may be that we simply lack an appropriate word for the divergence of capital prices from wages, productivity, and risk.
There is an undeniable societal impact of this divergence, individuals become less economically and socially mobile. They maintain a net debt rather than net asset count for longer, they may be either practically or perceived as locked out of societal progress.
My theory is that all money which could have been invested else where went to AI. It can end up either in investments paying off which will result in AI investors becoming even more rich (poor don't invest in the 1st place) and the rest of society poor or investments will have no returns and it will be wealth destruction on a grand scale and everyone will come poorer afterwards.
>only people >70 years old have ever really experienced it before,
Part of the observations from the HN reader on the current situation in Finland (believe everything he says, "canary" of Europe?):
>things really suck especially for fresh grads. There's fierce competition for jobs like cashier at supermarket, hundreds of applications for one position is normal. Lots of fresh grads with bachelor's or master's degree compete for those jobs too, since they can't find anything better.
In the USA this was one of the exact "unexpected" developments in the Nixon Recession that was like no-one had seen since the Great Depression. Except that depression there were not yet enough college grads in existence to contribute as a major statistic.
By the mid-70's I'll never forget the crowds vying for a single job opening at a gas station. Pumping gas and cleaning windshields when most stations had only converted half their pumps to self-service. Some with advanced degrees, it was not pretty. These were always minimum-wage jobs too, like supermarkets and fast-food.
When I started working there was a chemical plant within 25 miles where I could have gotten a job easily if I had graduated a few years earlier. Founded in the 1950's by one of my professors, it was actually pretty advanced. The placement office said they hadn't seen an opening in over a year. I was lucky to get a job at an appliance dealer because he liked my ability to program, but he never got a computer the whole time. Otherwise I wouldn't have got noticed, but the job was to prep merchandise for delivery so I was in the warehouse installing a lot of icemakers and doing minor repairs, plus rode along with a service operator one or two days a week to help when it was commercial refrigeration. Which I was learning, but also learned that I was kind of replacing an experienced repairman because they had let too many people go when things first got bad.
About a year later things got worse and he had to kick out the new people and we were gone.
I then began to get unemployment and the need had gotten so great that it must have been the first computerized institutional job boards for that reason. Slim pickings doesn't describe it. But you had to check in every week and apply to whatever you may be qualified for. I had gotten a cheap car (was riding bike to the appliance co) and was selling fruits and vegetables when a job came up at the plant. Not in a lab but out on the large reactor areas working with chemicals and taking readings. The posting had been badly mangled by the typist and it was not obvious it was a chemical job. You could still tell it was technical though. There were over a hundred applicants anyway, and not a realistic chance at all.
Months later I got a lead from my uncle that a lumber company near him needed somebody full time. This was about 35 miles away. All they would do is take my application without getting to talk to anybody, so it took longer to drive there. Still there were about a dozen people applying while I was there so it must have been hundreds of applications overall too.
Now I was already wearing a tie so I went back to the chemical plant and what really made the difference was that there were no new job openings so that time I was the only one showing up in a while. There was only one office building outside the gate, the manager was in and came up to see me but said right away they had no openings. He invited me for a quick tour of the labs and plant anyway, it was good getting inside the gate but like everybody else it was just optimism in the face of declining prosperity.
Surprisingly, he called me back a few days later and offered a part-time job, 4 days a week. He had talked to my professor, and I was a good student. I started out doing a lot of different things for different people, mainly for the analytical lab. In less than a year, they had me come in 5 days a few times when it got heavy and months later I was full-time.
I still wasn't getting twice the minimum wage, but I was so lucky.
After that I only sold produce on the weekends, and only seasonal things I picked myself like avocados and blueberries.
This is pretty ridiculous, just stupid enough for a bit of silly Friday watercooler conversation.
I have questions. How do facial expression, clothes, and hairstyle impact the model’s predictions? How about Facetune and insta filters? Would putting a clickbaity YouTube thumbnail at the top of my resume make me more employable?
This lines up with what I once heard “second hand” from faculty at a business school about publishing in academic business journals. It was something along the lines of being a bunch of dancing monkeys pumping out entertaining, to readers of HBR and such, content.
The word “smart” is doing a lot of work here. And I think both the author and a lot of commenters in this thread are equating it with IQ.
What about EQ? I would also consider people with high EQ to be smart. It is a different kind of smart, and provably one more correlated with happiness.
This article kind of grinds my gears. I feel like there is an unstated assumption that people in pharma R&D are idiots and haven’t thought of this stuff.
Pharma companies care very much about off target effects. Molecules get screened against tox targets, and a bad tox readout can be a death sentence for an entire program. And you need to look at the toxicity of major metabolites too.
One of the major value propositions of non small molecule modalities like biologics is specificity, and alternative metabolism pathways; no need to worry about the CYPs.
Another thing they fail to account for is volume of distribution. Does it matter if it hits some receptor only expressed in microglia if it can’t cross the blood brain barrier?
Also the reason why off targets for a lot of FDA approved drugs are unknown is because they were approved in the steampunk industrial era.
To me this whole article reads like an advertisement for a screening assay.
I work in drug discovery (like for real, I have a DC under my belt, not hypothetical AI protein generation blah blah) and had the opposite experience reading it. We understand so little about most drugs. Dialing out selectivity for a closely related protein was one of the most fun and eye opening experiences of my career.
Of course we've thought of all these things. But it's typically fragmented, and oftentimes out of scope. One of the hardest parts of any R&D project is honestly just doing a literature search to the point of exhaustion.
I side with you. The more you know, the more you discover what you don’t know.
Every attempt to consider the extremely complex dynamics of human biology as a pure state machine, like with Pascal, deterministic of your know all the factors, is simplification and can safely be rejected as hypotheses.
Hormons, age, sex, weight, food, aging, sun, environmental, epigenetic changes, body composition, activity level, infections, medication all play a role, even galenic.
Put it this way: even in Pascal (especially in Pascal) you generally work in source code. You don't try to read the object code, and if you do, you generally might try to decompile or disassemble it. What you don't do -unless you're desperate- is try to understand what the program is doing by means of directly reading the hexdump (let alone actually printing it out in binary!)
Now imagine someone has written a Compiler that compiles something much more sophisticated into Pascal (some 'fourth generation language' (4GL) ) . Now you'd be working in that 4GL, not in Pascal. Looking at the Pascal source code here would be less useful. Best to look at the 4GL code.
Biology is a bit like that. It's technically deterministic all the way down (until we reach quantum effects, at least). But trying to explain why Aunt Betty sneezed by looking at the orbital hybridization state of carbon atoms might be a wee bit unuseful at times. Better to just hand her a handkerchief.
(And even this rule has exceptions: Abstractions can be leaky!)
My main point of skepticism about repurposing is whether this is giving any of new and actionable information. It seems to be reliant on pre existing target annotations, and qualified targets already have molecules designed for them. Is the off-target effect strong enough to give you a superior molecule? Why not just start by picking a qualified target and committing to designing a better molecule without doing all the off target assay stuff first?
The list price is mostly a starting point for negotiations with PBMs and payers. Drugs are also often aggregated and bundled. So in a lot of cases is unclear what a drug actually costs.
reply