This book seems to be suffering from an identity crisis. It’s ostensibly aimed at “programmers” but the content is predominantly machine learning topics. What sort of general software engineer needs to know about GRUs and LLMs?!
Add that to the fact that it’s published by Packt - this is probably one to avoid.
I've been out of the loop for a number of years, but I believe there's a serious amount of ML being thrown at predicting time series, so this probably gives you an idea of how money is being made.
I've no idea what the current models look like. I imagine it was all RNNs but maybe transformers have taken over?
Privacy issues aside for a moment, this looks really good and is something I’d be interested in. I can imagine this being helpful for folks with ADHD or memory issues. I have found meeting transcription tools to be absolutely invaluable at work so to expand that to my wider life makes good sense.
That said I have to admit the monthly subscription price ($19-$29/mo) seems a little higher than I’d like for a wearable device, especially as the premium features appear functionally like a thin wrapper around an LLM and some search tools.
Indeed. I’d love that with glasses as well - having perfect photographic memory of everything I see or hear, augmented by an AI that can correlate events, would be a real superpower.
OTOH, I also watched enough Black Mirror to imagine the societal consequences of that.
And yes, the subscription model indicates they’ll know a lot more about you than the ad copy discloses.
If someone does one that’s fully local with absolutely no strings attached, I’d seriously consider it.
That pricing is for their other existing product/app that records and exposes digital activity locally in what looks like a fairly nice UI: https://www.rewind.ai/use-case/engineering
Very interesting company IMO although I haven't given it a shot.
> I've always been confused why salaried workers think they only need to work 40 hours a week.
Employment contracts in most developed countries will specify a) hours of business/office hours and b) an expected number of hours that the employee is to work. In many countries (including the UK for example [1]), there are even laws restricting the number of working hours by default.
This is great. Such a powerful visualisation for such a simple technique!
My only critique is that I wish the creator hadn't used JPG images - there are clear compression artifacts that appear towards the edge of the screen as one image transitions to another.
It's just because the image has to get so enlarged as it gets that big. You'd have to use massive images otherwise (I've made a few infinite zooms before)
> We believe, the SNN is the correct representation of the how the mind works. In fact, we have mathematical proof that our representation is the most efficient representation.
No details on the website, though. This is either very poor marketing, or (much more likely) snake oil.
I don't really know where to start after reading this - it reads like a collection of poorly-informed reasons that the author dislikes relational databases - almost all of which are entirely orthogonal to the concepts of relational algebra or modern database management systems.
For example, SQL injection can hardly be considered a problem with database systems alone - and the problem with using code injection as an example of how RDBMS are "so bad" is that code injection isn't inherent to databases - it's to do with executing unsanitised user input (which is a problem that should be handled in the software accepting user input).
Doing a little bit more digging, it appears that the author is peddling his own technology (the "contextual database") which is presumably the solution all these problems... but the product's website has such a scarcity of technical information that it appears to be almost entirely useless.
Humira really is a life-changing drug - but it is offensively, eye-wateringly expensive. Each auto-injector pen costs approx. £500; and I have to take one every 2 weeks. To put that in perspective, that's a cost of about £1000 per month - which is considerably more than my mortgage. My house costs less than my medical treatment.
Fortunately for me, I live in the UK, where the NHS is able to help absorb these costs. In the US, I guess they aren't so lucky - if I lived there, my Humira treatment would financially destroy me - and probably my entire family, too. It's morally reprehensible for a company to have access to such a life-changing treatment and then proceed to charge such disproportionately vast sums of money (certainly, far beyond the cost of R&D).
I suppose you could say that instead of crippling its patients physically (like their medical problems do), Humira cripples its patients financially - and I'm not sure that's a great deal better.
> In the US, I guess they aren't so lucky - if I lived there, my Humira treatment would financially destroy me - and probably my entire family, too.
What makes you say that?
Either private insurance would cover it, or Medicaid (for low income people) or they essentially give it away to uninsured people through a non-profit foundation. (AbbVie Patient Assistance Foundation)
I'm not especially familiar with the structure of the US health insurance system, but I think that the lower costs mentioned on the page you linked largely depends on either receipt of considerable amounts of state aid, or having pretty high-quality insurance.
Certainly, if you were to flick through various forums for those with Crohn's disease (e.g. [1]), you'd find many stories of people paying hundreds (or even thousands) of dollars for a month's Humira treatment.
Under current law (the ACA), the maximum annual payment would be $7,350 (on top of monthly insurance premiums).
That's for the lowest benefit, qualifying ACA plans and assuming the prescription is covered. The "cash" price is ~$4000 a month (https://www.goodrx.com/humira ) so even the lowest benefit plans are kicking in quite a bit of support (but probably also negotiating the price down quite a bit).
This is how the payment structure was explained to me when i started taking the drug. Ultimately i pay $5/mo, my insurance pays some amount, and Abbvie i guess pays the difference. I'm not sure what set of circumstances leads to people paying higher prices, given the info the commenter above linked to.
Yes as a chronic kidney disease patient I am glad I am in the UK id quite possibly be dead if I lived in the US.
BTW before my transplant a single one of my meds (phosphate binder) was around £500 a month god only knows what serverlamer costs in the states - let alone the cost of my transplant.
My girlfriend has lupus - diagnosed at eight. The lupus destroyed her kidney by her fourteenth birthday. She got a transplant at sixteen. Her transplant was covered by her mom's insurance with the rest of the costs falling onto our state's medicaid because she was a minor and her mom, an RN, was not well to do. Her medications were also covered by her mom's health insurance and picked up by our state's medicaid program.
She and her mom moved to another state when my girlfriend turned seventeen, so from that point it was her mom's insurance picking up the medication. She and I started dating when she turned eighteen. At that time, all of her drugs, including anti-rejection medications, were $5 per 90 day supply at the local pharmacy under her mom's insurance.
Her mom lost her job when she was 21, so my girlfriend lost her insurance and was unable to afford COBRA. My girlfriend did not qualify for medicaid in this new state because she was not a minor, pregnant, or considered disabled by the federal government despite not being able to work with the lupus. At this time, our best option for her medications were discount mail order pharmacies, which cost $1000 per 30 day supply, which we could barely afford with my computer programmer salary, as I could not add her to my insurance due to us not being married.
For awhile, she was on PCIP and then a plan made available on the ACA exchange. Her premiums were $400 per month, and that made her prescriptions $20 per 90 day supply. So it brought the costs from $1000 per 30 day supply to about $450 per month, counting insurance. We managed to make this work for a few months before I was laid off from my programming gig and lost my insurance.
We moved back to the state she moved from due to a job offer. She still did not qualify for medicaid, but my new job allowed me to add someone to my insurance if we were married or domestic partners. So we got a domestic partnership, and her drugs now again cost $5 per 90 day supply.
So I've been with someone on both sides in America - good insurance, no insurance, basic insurance, and then good insurance again. We really need to find a better system. We've survived by luck and really tight financial planning.
> My girlfriend did not qualify for medicaid in this new state because she was not a minor, pregnant, or considered disabled by the federal government despite not being able to work with the lupus.
What state was this in? In Maryland, a single individual is Medicaid eligible up to an income of $16,600/year. There are 18 states where non-disabled, non-pregnant adults are not covered by Medicaid. That's not an "America" problem, it's a problem with those states and their voters.
I'd rather not say which exact states, but she was non-pregnant and not considered disabled by the state or the federal government despite applying for it, having the statements of her doctors, and having had a kidney transplant, chronic pain that prevented sitting and standing, blood clots, and lupus. So in the eyes of the state, despite having no income, she was not eligible for medicaid. They did tell her that if her kidney started to fail again then they _might_ consider her disabled, but she shouldn't expect another transplant because you typically only get one.
She was recently determined medically disabled by the federal government and now just needs to get through the financial eligibility steps.
I think that's a problem with the USA the fact that your rights and statutory benefits vary by where you live would be an anathema in any other first world country.
I had some swelling in my legs and chest pains recently and had to make a trip to the ER. It was very difficult to weigh the threat of the symptoms against the potential cost of an ambulance ride and emergency care, and it occurred to me at the time how ridiculous it would be if I were sitting around performing cost benefit calculations before calling the police or the fire department. I've really been thinking hard about whether I want to continue living in the US as I age.
I wonder what’s the reason for the downvotes. This was a purely factual comment directly addressing the parent’s point of “quite possibly [being] dead if [he or she] lived in the US”.
Kidney failure is in fact the best example of the
US government stepping in to provide universal coverage for a life saving treatment (even though it requires copays, but there are also patient assistance programs to help with that; see below).
A different question is that the system has many issues, see for example [0] and the recent figth between private insurers and dialysis clinics due to the assistance provided by charities which are funded by the dialysis companies that help patients to pay for commercial insurance (reimbursement from private payors is much higher, even though most patients are on Medicare they are not profitable).
the price the NHS will be able to negotiate will be a lot lower too
a natural result of the having an organisation that represents 65 million, backed up by leading researchers, instead of each individual completely on their own
Actually, I don't think that's quite correct. The specialists that prescribed me the Humira treatment were quite keen to impress upon me the cost of the drug - i.e. that it "costs the NHS over £400 per pen" (and this was back in 2015). Of course, this is purely anecdotal but I'm inclined to believe the source as they've no motivation to mislead me.
I take Remicaid, which has technically been on the market longer than Humira and still requires me to go in to an Infusion Center to get hooked up to an IV for 2hrs every 8 weeks. The ‘full price’ cost of my treatment: $16,000 just for the Remicaid.
They have a co-payment plan that, for the most needy, would effectively make the drug free. For me, my insurance gets billed a few thousand and I would have a $250 copay, which they pay.
It makes me wonder though, has anyone ever paid the full price, or is it just astronomically high to maximize the ‘charitable’ deduction side of their business?
I haven't had to pay full price. But, last year I had to pay 20% of the cost. It gets out of control fast since I have to take Remicade every 4 weeks. I was on the hook for about $6000. My bill was cut down because I qualified for charitable funds from my hospital.
It's shocking to see the bill before insurance pays it's share. I was getting bills for $25k. !!!
I can't speak for the US, but I've been told by the clinical specialists at my local hospital that Humira cost £400+ per pen (this was back in 2015). I don't think that AbbVie can offset any of that price as a charitable donation, so I guess that price must be what they actually charge.
Note that the list price in the US is way higher than that (over $2000). But list prices are not actual prices (the average discount may be around 25% for Humira).
I think your mortgage/rent should be lower instead of the drug. Why do people think giving $ to landlords is better than scientists/doctors/researchers/business ?
Bearing in mind that accommodation is quite frequently the single largest outgoing in many people's finances (i.e. commonly approx. 30% of income), do you really think that the cost of medical treatment should exceed this? Do you really think that medical treatment should be an individual's single largest outgoing cost?
As for giving money to scientists/doctors/researchers, that isn't really the case either. See [1] - the cost of prescription medication appears to be largely determined by "what the market will bear", rather than how much a drug costs to research and develop.
Perhaps the NHS shouldn't - not paying for Humira would essentially obliterate AbbVie's UK market, which would make a considerable dent in their profits... I guess that would probably make them re-think their pricing structure!
However, presumably the treatment is "worth" the NHS paying for it - even if it is massively overpriced. I can't really pass comment on their medication choices.
I think medical treatment SHOULD be the largest cost. Do you really think something else is more important ? At least non-sick people will have money this way.
I understand, what the market will bear. But that's for everything ? (market bearing rents in sf)
Still, drug companies are more prone to invest their $$ in drugs compared to landlords, don't you agree ?
Of course I'm not ok with spending 30% on rent + 30% on meds. But it's more insane that we do on rent, I think (or move house and commute 2 hours, waisting your life)
I'm trying to fathom your point. Regardless of how much rent or mortgages cost, do you think it's acceptable or fair for individuals with chronic illnesses to bear the cost of a treatment which would cost them more than their rent (i.e. an additional cost which healthy people wouldn't have to pay?). That sounds like a pretty bad deal if you ask me - not only would one have to deal with a chronic disease, one would also have to budget for a (very considerable) additional expenditure which healthy individuals wouldn't have. At the end of the day, that is what makes the price of Humira unfair and immoral - its ability to severely financially disadvantage its patients (by being priced disproportionately compared to the cost of R&D / production).
Now, if you're talking about spreading the cost of all medical care across the entirety of the working-age population (essentially, national insurance) - that's a more sensible proposition. In fact, that's (broadly speaking, of course) what the NHS does in the UK - and the system works quite well.
> It's morally reprehensible for a company to have access to such a life-changing treatment and then proceed to charge such disproportionately vast sums of money (certainly, far beyond the cost of R&D).
I don’t really understand. If there is a moral obligation to help cure other peoples’ diseases, shouldn’t it be even more morally reprehensible to spend billions on things like TV serieses, cell phones, and social media websites, which could’ve been used to cure diseases instead? By your logic, isn’t Apple like the most evil company in the world, because they spend billions of dollars that could be used for drug research on making iPhones instead? Why is it somehow worse to make a profit curing people, than to make a profit not curing people?
Because the profit made by "curing people" is actually done by keeping others from curing people, via aggressive patenting, FUD over generics, lobbying against legislation that would make prices reason able...
There's also the fact that public health is not a simple economic market, where customers can simply not buy your product. In a lot of cases, it is a life or death situation, and real human have to be treated, so it should be considered a public service.
> Because the profit made by "curing people" is actually done by keeping others from curing people, via aggressive patenting
It’s not like the treatment was just out there until AbbVie monopolized it. Chron’s disease was medically characterized in 1932. Anyone could’ve come along and developed this treatment and gave it away for free or cheaper.
> lobbying against legislation that would make prices reason able...
That just begs the question of what’s “reasonable.” What single payer does is create a monopsony (single buyer situation). Obviously it lowers prices, that’s what monopsonies do. But like it’s counterpart, the monopoly, it’s economically inefficient.
Can you propose another way of financing the cost of drug discovery, research and testing for safety? The current system works reasonably well but if all you’ve got is outrage I would prefer at least outrage and a plan.
Read the findings from [1] - it appears that in a very considerable number of cases, there's actually a disconnect between the cost to research, develop, and produce a drug - and the price that's charged to customers (or insurance). The conclusion is that the price is mostly determined by "what the market will bear", rather than much else. This suggests (to me, at least) that in cases of a particularly expensive drug like Humira, a large proportion of the profits are not reinvested into R&D.
I suggest reading a microeconomics textbook. In the absence of price controls price will be set such that supply equals demand. It does not surprise me in the slightest that a drug that people really, really need and that is still under patent protection is very, very expensive.
Prices are what makes supply equal demand.
Blockbuster successes, massive profit makers are what pays for all the many, many failures of drug development. There’s quite a bit of research on the economics of drug development but the thing that I find the most convincing is that the big drug companies aren’t exceptionally profitable. They’re pretty much in line with what you’d expect of multinationals of their size, 10% or so profit on revenue per year. Scientific publishers like the accursed Elsevier earn supernormal profits, 33%.
It appears that net profit (for AbbVie, at least) tends to sit quite comfortably between 20-30% most years - 24.27% last year, and hitting almost Elsevier-like levels (~32%) in 2009 [1]. That smells like profiteering to me.
Add to that their particularly hostile approach to patent applications (100+ on Humira alone), and to me it appears that AbbVie perhaps has a vested interest in stifling innovation and charging large amounts of money for their products "because they can".
Call me when AbbVie is listed with Roche, Novartis, Pfizer or GlaxoSmithKline as one of the big pharmaceutical multinationals. A relatively small company having a blockbuster success and returning huge dividends to shareholders is why people are willing to invest in biotech/pharmaceutical startups at all. Every biotech startup formed around one drug that gets serious funding burns at minimum a decade of man years and at the beginning most of those man years are going to be researchers with doctorates. Absolutely nothing gets to market with that little investment and most of these companies fail.
The point about the pharmaceutical sector’s profit margins is about the sector more than it is individual companies. Minnows that make supernormal profits expand and as they expand their profit margins decline because they use up the most profitable opportunities and move on to less profitable ones until there’s nothing left. This last bit never actually happens because things change fast and general equilibrium is never reached but that’s the tendency. Large profit margins attract competitors and in the meantime allow those enjoying them to grow.
Re: “Profiteering”
According to the theory of neoclassical economics, anti-price gouging laws prevent allocative efficiency. Allocative efficiency refers to when prices function properly, markets tend to allocate resources to their most valued uses. In turn those who value the good the most (and not just the wealthiest) will be willing to pay a higher price than those who do not value the good as much.
That’s about the same profit margin as Google and Facebook and Apple. Are they “profiteering?” Saying that the “reasonable profit” for drug companies should be lower than tech companies, say 15%, is economically equivalent to saying we should impose a 50% tax on drug company profits. Which would be nuts. You tax products you want less of (cigarettes, sugary drinks) not ones you want more of.
I’m not asserting that you’re drawing the line in the wrong place. I’m asserting that you’ve got the arrows backwards. Curing people and charging too much is >= not curing people at all.
Humira doesn't cure diseases; it only treats them. Many drug companies have stopped funding cures because biologics like humira are just too profitable. Why charge a one time fee when you could bill someone for life?
That’s oft-repeated but makes no sense. If you develop and patent a treatment, then that necessarily means that the cure—if it exists—is out there and unpatanted, for some other company to develop and sell. There is an economic incentive for them to do so, because a cure is obviously more valuable than a treatment. It would be very weird if you could somehow make a treatment more valuable than a cure by changing how you charge for it. By that logic, companies would never sell cars, only lease them.
Counterpoint: in the US my employer-based insurance paid for it, while in the UK the NHS refused to cover it. I was also covered in France and Germany, so the UK is the odd one here.
You're implicitly making the assumption that an algorithm is unbiased and less likely to be fallible - and unlikely to be susceptible to bias or manipulation (at least, compared to a human). There is a considerable amount of academic and popular literature showing that these assumptions are false, at least at a technical level.
The problem of vetting YouTube videos isn't necessarily something that can be solved by an algorithm, at the end of the day. Lots of the simple cases, sure - but such an algorithm won't work on even the most moderately challenging content. After all, moderating complex content such as video doesn't just require a well-trained neural network to recognise objects portrayed in images and sounds in audio... deciding whether a video is appropriate or not for monetisation is something that requires an explicit understanding of the social and cultural context which the video is going to be viewed in - which is both highly subjective and difficult to express in an algorithm.
NB;
- Neural networks (and other AI techniques) can be tricked and manipulated (see [0,1]; examples of adversarial images)
- Machine learning algorithms can induce or confirm bias through inadequate training data and poor assumptions. (see [2]; a book about bias in machine learning algorithms)
> You're implicitly making the assumption that an algorithm is unbiased and less likely to be fallible - and unlikely to be susceptible to bias or manipulation (at least, compared to a human). There is a considerable amount of academic and popular literature showing that these assumptions are false, at least at a technical level.
No, I'm arguing that algorithms are transparent, which means that they can be adjusted.
> The problem of vetting YouTube videos isn't necessarily something that can be solved by an algorithm, at the end of the day. Lots of the simple cases, sure - but such an algorithm won't work on even the most moderately challenging content.
of course they would here:
while (!decision_pass_the_spot_check()) { tweak_algorithm() }
> deciding whether a video is appropriate or not for monetisation is something that requires an explicit understanding of the social and cultural context which the video is going to be viewed in - which is both highly subjective and difficult to express in an algorithm.
Certain cultures think it is acceptable to stone women. In their societal context it is OK. We do not, however, care that it is OK in their societal context. So we make sure that in our algorithm stoning women ==> demonetized. Now we move onto the next problem. Did someone flag a video that we processed that had women stoned that was not demonetized? We open a ticket with the people responsible for that model, with the video that was not demonetized and have them figure out why it did not happen.
That's how you solve a complex problem. You break them into simpler ones and when you see an issue you address it. Claiming that the problem is just too complex for algorithms is avoidance.
Your argument that algorithms are transparent is equally flawed. There is an entire field of research dedicated to introspection of layers in convolutional neural networks! A system that is complex and nuanced enough to be able to deal with a problem such as auto-moderating YouTube videos will be phenomenally difficult to inspect (and nearly impossible to "adjust" in the way that you seem to be suggesting).
To clarify; I am not saying that it will never be possible to solve this sort of problem with an algorithm - just that doing so would require solving the small inconvenience of general artificial intelligence first... :-).
It certainly isn't beyond the realm of possibility in the (very distant) future, but to suggest that it's possible right now (and that it would be better than a human!) is to over-egg the pudding somewhat.
It is not flawed at all. If you cannot provide a validation for your algorithm decisions when challenged then you algorithm belongs in a research lab or on your laptop, not in production.
>It is not flawed at all. If you cannot provide a validation for your algorithm decisions when challenged then you algorithm belongs in a research lab or on your laptop, not in production.
Alright, then you cannot use an algorithm for image or video recognition. So how do you solve this problem?
Okay, then your argument is valid... in a different layer of reality. Meanwhile, everyone and their dog are rolling out blackbox neural net classifiers to production
Add that to the fact that it’s published by Packt - this is probably one to avoid.