Strawman. There's very little pressure, if any, to drink all the time. On the other hand, businesses are pressing employees to constantly use AI tools. The sad fact is that in many cases, they do result in faster work, but at the same time they carry so many negative effects for the individuals.
Because bonds involve interest. Per Summa Theologica:
> To take usury for money lent is unjust in itself, because this is to sell what does not exist, and this evidently leads to inequality which is contrary to justice.
…Aquinas expands the analysis but it is relatively straightforward: all interest is usury.
Personally, I find it helpful to imagine two hypothetical persons representing the entire economy, one the creditor who is lending and two the borrower who is taking on the loan. In this ultra simple closed model with a fixed quantity of money, the former is in effect asking for more units of money than actually exist in the whole system. When the loan comes due the borrower owes a sum that cannot be paid in full from the circulating medium itself. Settlement then requires either default, the creditor forgiving the excess, or the transfer of real goods and property to make up the difference. Scaled up, that same pressure (the continuous generation of monetary claims that exceed the existing stock of money) is what I suspect drives a good deal of the subtle and overt strain on families and communities that people so often complain of in the West and in modern growth-oriented capital societies.
This definition of usury differs from the modern loophole-definition: that interest bearing loans are only usury when the rates cross some nebulous abusive threshold. In the above Thomistic interpretation, all interest is socially problematic and disfavored. Judaism holds to a similar prohibition on interest when loans are made between Jews. Islam likewise prohibit usury even more broadly. Despite the injunction against usury in the Middle Ages Christendom and the enduring prohibitions of usury in other faiths, there are many modern Catholics and Protestants who will favor the modern interpretation over Thomas’ understanding; I’m just not one of them.
>in effect asking for more units of money than actually exist in the whole system.
Depends on how you define the 'whole system'. If I borrow $100 and make $110, the latter didn't appear out of nowhere. The lender, too, could have turned that $100 into $110.
Why shouldn't they be compensated for that opportunity cost?
If I understand right, it is allowed to invest like a partnership, where you make $110 together out of your $100 and your partner's effort. But you must be exposed to the downside of failure just like your partner.
The opportunity cost point is fair enough if one is thinking in terms of two concrete individuals. The illustration I offered was meant at a more abstract level, the two persons standing in for the creditor side and the debtor side of a closed economy taken as wholes. In that framing the issue is not whether a particular borrower can put the money to productive use (clearly he can), but that the system as a whole is being asked to generate more units of the circulating medium than currently exist within it. Even when real value is created, the monetary claim still exceeds the monetary stock. Settlement then requires continuous expansion of the money supply, continuous transfer of existing assets toward creditors, or periodic default. That structural pressure is what I was trying to get at.
I suspect the deeper difficulty is the “bond” in bonds themselves, the ongoing compulsion that interest introduces. Once interest is attached the debtor is under continuous obligation to produce additional claims simply to keep the accounts from breaking. Traditional writers on the Christian and Islamic sides generally preferred arrangements that avoided this continuous pressure. A pure discount (as with discounted Treasury bills and similar instruments) prices the time element once, up front: the creditor advances a smaller sum and later receives the larger face amount. The cost is paid at the beginning rather than levied as a recurring claim that must be met out of future circulation. In that sense the time value is acknowledged without the mechanism that forces the system to keep generating more monetary units than presently exist.
This still seems off to me. If I were a potential creditor in such a closed loop system, and I was told I absolutely could not charge interest due to these monetary supply constraints, I would either just stop lending entirely or I would demand something that isn't strictly denominated in money to make the risk I'm taking on, etc worth my capital outlay.
But then eventually, if the system were sufficiently complex, I'd probably tire of whatever complicated barter system we have already going on, and then it's likely some third party would step in offering something that's totally not money, dude, trust me, it's just like a handy clearinghouse of IOUs for people engaged in the trade of these non-monetary favors for favors...
Some people who hold or offer such IOUs might then take the bold step of calling them non-exclusive, as in I will mow the lawn of whoever happens to have my "one lawn mowed" voucher, I just happened to originally give it to this first guy, I have no idea what he did with it after that... Other people realize this "non exclusivity" deal actually makes the voucher strictly more valuable, you can do more things with it than you could otherwise... You see where I'm going with this. It's not passing my sniff test.
But, there is no fixed quantity of money in modern finance. It's created every someone or some business takes out a loan from a bank, and every time the government spends money. It's destroyed when the loans are repaid or taxes are paid.
Greetings, fellow Thomist. You are correct that usury is any amount of interest, and attempts to pretend otherwise are sophistical, but I suggest you may want to look again at corporate bonds and other non-recourse loans -- that is, loan where collateral is limited to specified asset(s). Look at Zippy Catholic's writing on the subject -- he was a finance dude, and a very rich self-made man. A non-recourse loan is more like taking an ownership share in something, and then renting it back. It's unfortunate that we use the same words (loan, interest, debt) for both full-recourse and non-recourse contracts, because they're entirely different things.
I agree that charging interest on a full-recourse loan is a wicked and disgusting thing to do to one's fellow man, and I'd say it's in the same genus as slavery. Usury is to fraud what robbery is to larceny. It's also interesting that the markets where usury is most prevalent (housing, college fees) are the ones that have seen the most insane price increases.
In many historical societies, religious prohibitions on usury meant the charging of interest of any kind.
Jump in a time machine to 1515 and ask Martin Luther, or to 1260 and ask Thomas Aquinas, they'd tell you it's sinful.
And in the present age, a fair number of Islamic folk consider interest against their religion's rules. So there's a Halal finance industry where, for example, you can get a "murabahah contract" where the bank buys a house, then sells the house to you at a higher price, while allowing you to pay them in monthly instalments.
I love when religions have rule lawyers like this. It readily discredits the religion. As if their all powerful god can be fooled by fancy paperwork or legal loopholes.
They’re not trying to fool God, they’re trying to fool you into going along with it. They don’t care what God thinks and may not even believe in Him at all, but unless they can convince you of the loophole they’re stuck with the rules themselves.
The bit that isn't rules-lawyered away is that the risk is shared. For the deal to be compliant with the religious law, the lender must accept the same risk as the borrower, equally.
So I guess in this case if the house burns down and the insurance only pays 50% of the agreed value then the lender only receives 50% of their agreed repayment.
"Thou shalt not lend upon interest to thy brother: interest of money, interest of victuals, interest of any thing that is lent upon interest" (Deut 23.20 JPS Tanakh).
It's concretely different. If the house becomes worthless, the "borrower" can walk away from the contract, owe nothing, and the bank keeps the house. The bank had better consider the value of the house, not just the ability of the "borrowed" to pay, when issuing this contract.
I absolutely agree, although even that doesn't solve the root problem. The underlying LLM architecture is fundamentally insecure as it doesn't separate between instructions and pure content to read/operate on.
I wonder if it'd be possible to train an LLM with such architecture: one input for the instructions/conversation and one "data-only" input. Training would ensure that the latter isn't interpreted as instructions, although I'm not knowledgeable enough to understand if that's even theoretically possible: even if the inputs are initially separate, they eventually mix in the neural network. However, I imagine that training could be done with massive amounts of prompt injections in the "data-only" input to penalize execution of those instructions.
I think there are two distinct attack types for LLMs. Jailbreaking is what most people think of, and consists of structureing a prompt so the LLM does what the prompt says, even if it had prior context saying not to.
The other type of attack would be what I would call "induced hallucinations", where the attacker crafts data not to get the LLM to do anything the data says, but to do what the attacker wants.
This is a common attack to demonstrate on neural network based image classifiers. Start with a properly classified image, and a desired incorrect classification. Then, introduce visually imperceptible noise until the classifier reports it as your target classification. There is no data/instruction confusion here: it is all data.
The core problem is that neural networks are fairly linear (which is what makes it possible to construct efficient hardware for them). They are, of course, not actually linear functions, but close enough to make linear algebra based attacks feasible.
It is probably better to think of this sort of attack in term of crypto analysis, which frequently exploits linearity in cryptosystems.
The depth of LLM networks make this sort of attack difficult; but I don't see any reason to think you can add enough layers to make it impossible. Particularly given that there is other research showing structure across layers, with groupings of layers having identifiable functionality. This means it is probably possible to reason about attacking individual layers like an onion.
This problem isn't really unique to AI either. Human written code has a tendency to be vulnerable to a similar attack, where maliciously crafted data can exploit the processor to do anything (e.g buffer overflow into arbitrary code execution).
However, you may immediately see how using same input space essentially relies on the model itself to do the judgement which can't be ultimately trusted
> one input for the instructions/conversation and one "data-only" input
We learned so many years ago that separating code and data was important for security. It's such a huge step backwards that it's been tossed in the garbage.
I find Gemini is outstanding at reasoning (all topics) and architecture (software/system design). On the other hand, Gemini CLI sucks and so I end up using Claude Code and Codex CLI for agentic work.
However, I heavily use Gemini in my daily work and I think it has its own place. Ultimately, I don't see the point of choosing the one "best" model for everything, but I'd rather use what's best for any given task.
You don't need to propagate it, you just need to show the gradient of the current position alongside with the classical evaluation, to give more context to the viewers.
Why would they be? Cursor took an existing editor and added some AI features on top of it. Features that are enabled by a third party API with some good prompts, something easily replicable by any editor company. Current LLMs are a commodity.
You could be right, but I suspect that you're underestimating the degree to which GPT has become the Kleenex of the LLM space in the consumer zeitgeist.
Based on all of the behaviour psychology books I've read, Claude would have to introduce a model that is 10x better and 10x cheaper - or something so radically different that it registers as an entirely new thing - for it to hit the radar outside of the tech world.
I encourage you to sample the folks in your life that don't work in tech. See if any of them have ever even heard of Claude.
I don’t think people outside of tech hearing about OpenAI more than Claude is really indicative of much. Ask those same people how much they use an LLM and it’s often rare-to-never.
Also, in what way has OpenAI become the Kleenex of the LLM space? Anthropic, Google, Facebook have no gpts, nobody “gpts” something, nobody uses that companies “gpt”.
I would say perhaps OpenAI has become the Napster, MySpace, or Facebook of the LLM space. Time will tell how long they keep that title
