13,709 karma · joined March 30, 2021
Mastery of arithmetic involves memorizing lots of heuristic rules that aren’t particularly insightful about, or are outright conflicting with, the structure of math, and learning the structure of math first slows your ability to do arithmetic.
I don’t think any school hands first graders calculators and asks them to never learn arithmetic. But I see no need once you’re doing algebra to be focusing your time on long division or multiplication. These are irrelevant distractions from algebra, time spent and points lost on arithmetic details is a waste of cognition and doesn’t help. This becomes increasingly true through calculus, linear algebra, until you get to real math courses where a calculator is pointless because arithmetic doesn’t even show up.
The fact is a lot of people have trust issues, anxieties, social issues. Some are paranoid, suffer extreme neuroses, or panic in a social situation. Some have felt betrayed by professionals seeking to help them, putting them into inpatient care against their wishes, or medicating them with debilitating medications to sedate them, which is common for people experiencing mania. That’s not a personal failure, that’s just personal issue. And someone seeking mental help is having personal issues more often than not, unless they’re just seeking life coaching.
Ascribing these as fundamental personal failures, and therefore they deserve their personal hells with no relief unless they comply with your view of how people must be, is cruel.
The sycophantic behavior you’re outlining is a result of reinforcement learning and alignment, not a fundamental property of generative language models.
Fine, let’s have socialized delivery of mental health care. Let’s go back in time 20 years and establish a pipeline of capable providers that doesn’t exist. Let’s do those things if we can. But I don’t see it happening in my life time.
I do however see an opportunity to take the technologies presented to use in the immediate present and providing a mechanism to bring help to people who need it today with good efficacy where it’s appropriate. Is it OpenAI that brings that? I doubt it - they’re selling shovels, not digging for gold. Is it a general model that provides direct care? Doubtful IMO. Is it a chatbot interface ? I doubt that too. But I think if we are creative and thoughtful we can use these technologies to address huge swaths of the care providing problem, which is more than talk therapy. And we can do it today, and don’t have to wait for MAGA/MAHA to accept reality on realities terms.
Does that mean in 2026 that all ICE use cases are replaced? No, and with the aggressively anti science US administration it’s been set back by at least 8 years - the time it will take to rebuild momentum lost, research discarded, investment displaced. But there’s no reason beyond our lack of imagination to believe ICE will be on the road in 50 years in any use case.
Also, it’s a different mode of delivery. Some people work well remotely, others better in person. Likewise some people open up more easily to a machine than a person. One doesn’t replace the other.
Further, LLMs can be very useful for diagnostic work in addition to a psychologist to help accurately classify mental and personality disorders. This is my fathers speciality and he finds modern LLMs to be fairly adept at classifying.
A lot of psychologist work in a clinical setting is case notes and an LLM can be quite skilled at producing standard expected case note structures from clinical notes, improving the time spent by the professional in care vs paperwork.
All of these are opportunities for use of LLMs as specialist models, agentic frameworks around general models, guard rails, delivery mechanisms, etc. I wouldn’t look at the world being OpenAI and Anthropic, but as model providers as providing screws and bolts and it’s up to us to build the products. LLMS are trained with an enormous corpus of witting on mental illness, but also writings by the mentally ill, care professionals, and all sorts of other materials. By eliciting all these dimensions the applications to mental health are enormous - it is a giant learned model of human thought and experience, not just of factual information, but the entire complex.
Here’s an idea I’ve talked to my dad about - how about using a model set trained on the writing of many people of specific mental illnesses. Then you can interact with them and they will produce responses likely of someone with that illness. This can be used to train clinicians, but also to build better inventories and diagnostics since you know the classification of the responses to some reasonable degree. This should accelerate research into mental illnesses.
The possibilities are vast, limited by our imagination and biases.
My parents are clinical psychologists with 40 years experience, and they both agree LLMs could be a great boon to many many unserved people if made well. They think this is a wonderful avenue of research. But agree in the moment it’s not ready. I think the ethics of allowing people to suffer due to economics while ignoring a technology that can help them is worse, no? Telling them to seek a health professional when none are available is worse ?
Plus they last basically forever and require almost no maintenance. I like being able to never change oil, tune up, belts, whatever every few months. Every few years I get the cabin filter changed and that’s it. In 10 years the battery might be reduced by 20%, and I might replace the batteries around 15-20 years out, but I expect to literally drive it into the ground. (Mine is a Tesla with unlimited super charging for life, and I live next to a super charger)
I have family overseas and will likely buy them one of these EV Camrys. If they’re good, I might get one for my daughter. At $30k it’s fairly affordable.
The point of a grep is for adhoc analysis of data that isn’t in a database. At best you could use an inverted index, which is extremely expensive to create and maintain, but if you don’t know a priori what you’re going to search for, or even if an inverted index helps with your query, it’s pointless and expensive to index literally everything you could ever use to save some time on a tool call.
The point isn’t to find a way to make searching more efficient. It’s that LLMs are becoming very efficient, and a measure is relative to grep. Making grep and searching more efficient has been a task in CS for decades. LLMs is relatively new beyond a very small crowd. It’s also not a goal to make LLMs better at grep than grep, but as a relative reference point that inference is getting cheap and fast and has opportunity to get faster far beyond grep simply due to unclaimed optimizations more available for the newer technology.
An LLM can certainly be cheaper than grep, because it’s an approximation, while a grep is deterministic and must examine every byte in what can be a relatively complex state machine for a regex based grep. There are other scales to consider like the scale of your local hardware vs the highly multitenant and high end hardware of the hyper scale inference providers.
There are already high volume models for coding inference where the reasoning time is crazy low and cheap per token where it can build reasonably simple software so blindingly fast it isn’t implausible the bottleneck is the latency in tools and networks. I find them hard to use at times because I don’t have time to think through the next turn by the time it’s done.
Regardless I wouldn’t be surprised to see a world where tokens are so cheap it’s not worth metering them but charging licensing feels with meter tiers at the far horizons to prevent abuse, charge outliers. Subscription models already set this stage well.
The other side to consider is bountiful capacity will also drive tokens to near zero price. The data center build out is barely underway and as it materializes, as hardware efficiencies improve, as techniques and model science and technology improves, harnesses, methodologies , etc improve, the economics flip from load shedding to trying to keep the data centers utilized. The economics lead to the world where tokens are not a unit of measurement for cost for anyone other than the inference providers to manage their utilization.
https://gizmodo.com/trump-says-hes-changing-artificial-intel...
Overall it was not bad. The cred vault is useful. But I’m not giving meta any information that I wouldn’t otherwise be fine with disseminating and I feel the burden the school puts on us vs the information loss is worth the trade.
Thanks! First meaningful use of an OpenClaw like system.
Not only are political officers making science decisions, but nasa lost its way in the depths of over engineered missions due to requiring a senator to be bribed by the quantity of home stage work.
Homework isn’t assigned in class any more. It’s posted on dozens of inscrutable websites in weird UX’s that are impossible for any single human to understand. Children are mystified, parents confused, teachers at their wits end.
A few times a week children miss some website or queue from some system, teachers can’t get things to share properly, this needs tending, they need to photograph this and upload that to this place as well as that but not the other one.
It’s truly remarkably astounding. I have serious doubts frontier models can navigate it either honestly. But this is in fact where I see the most chance of using these tools.
“Computers at school” is the worst idea ever, and no one who hasn’t experienced the clusterfuck of middle school in 2026 can understand it. It makes the world of the movie Brazil seem utopian.
Does anyone have any meaningful use cases that have actually improved your life in some measurable way?
I’ve tried hard and beyond filling out some tedious forms in about the same time and energy as just doing it myself, I’ve not personally. I’ve tried hard. So I figure it’s probably my lack of creativity given the prevalence of fans. But every time I ask, I get no meaningful responses. So it makes me wonder if all the fans are just product managers on those products?
Self run infrastructure won’t have this cost but you have to manage the capacity and rollouts yourself, at which point it’s more obvious what’s happening, but the effects will be the same. The not knowing makes it harder, but also harder to plan your own work around.
Fable seems to be following the same enshittification arc of other Anthropic models.
Generally OpenAI seems to be taking the opposite approach with an increasing improvement over time. As sad as I feel to say this, open ai seems to have the right strategy. Making your product worse over time rarely plays well with customers. At this point it feel often hard to justify using Anthropic for anything. I generally like Anthropic better as a company and they really had the initiative and advantage and customer good will, then proceeded to squander it faster than a cigarette company or the Sacklers could have.
I’d also note you have to be at a company that even offers a 401k. Only 45% of Americans have access to a 401k, or have their own IRA. This isn’t a safety net - I’d like to see you do a high wire with only 45% of the area of your walk covered with a net, and of that the median segment doesn’t have the strength to hold your weight. It’s a political gambit to fool people into voting against their best interests.
Most people having had a rental at some point understand the concepts a lot better than bond trading and investing writ large. They also can’t convert their basis in their home appreciation into their 401k, and my point was on basis rolling of property capital gains and how to convert your family home appreciation into a tax deferred vehicle with cashflow.
Personally, I agree with you; I would do a self managed annuity structure and hire an advisor for my later life when I may not be able to manage it. However, I’m just explaining the typical thinking. And the capital appreciation basis roll is a real problem for retirees whose home appreciated over their lives.
I think you misunderstand what happened in the 80’s - pensions were raided and they were stolen. In the 90’s the defense was to not offer them any more. There do exist some pensions, mostly governmental, and some rare corporate pensions that escaped raiding, but the survival of private pensions wasn’t great even for those that had them. Social security is below poverty rates, and is not a retirement plan but supplemental. It was designed with the private pension system existing in mind. I’m fine with expanding social security by lifting income caps and other fixes, including a sovereign wealth fund. The status quo isn’t tenable.