morelandjs

I like how you have to be an accredited investor to get involved with riskier investment instruments but you can be a total moron with no money and lose whatever you have left betting on sports and prediction markets.

Btw it’s not just those two, add iGaming into the mix which is cell phone gambling glammed up in Candy Crush form. Leeches, all of them.

In many cases, you can partition data at rest (eg parquet) on the user access key, then load it once into an agent container at runtime. In this way, one authentication pass provides access to all data in a sandbox, versus repeatedly authenticating against MCP calls. It also just reduces the connective tissue between the harness and the data. There are obviously drawbacks in that you aren’t leveraging real time APIs, but it works well for most business use cases.

Was ready to write something snarky because this is essentially RAG, but I think the author is getting at some subtle details which are seemingly important.

- memory systems are a specific type of knowledge base where you generate all the documents. You might as well generate them to be less than your embedding token limit to obviate the need for chunking.

- embedding models are getting better and are no longer just semantic averaging.

- small models are getting dirt cheap, making parallel reads cost manageable

What they describe is sort of the simplest architecture that takes advantage of these observations. I believe them when they say it works well.

I do suspect though that things like keyword lookup will completely fail if every memory is just a vector. Hence why something like Typesense hybrid search can still be useful.

Worth mentioning that controlling for confounders is hard, a reasonable prior is the null effect, and any time a researcher makes a mistake it produces a publication claiming a novel effect. You observe a set of publications concentrated with false positives as a form of selection bias. I wonder how many of the causal studies are reproduced by independent teams using independent datasets.

Plausible scenario. Individuals predisposed to Alzheimer’s experience different mental sharpness from birth and this makes them enjoy taxi driving less (more intense than long-haul trucking), and so they pursue taxi driving as a career at lower rates. Under this scenario, driving has no effect, it just induces a selection bias.

Maybe confidence is a loaded term, but LLMs do have a sense of procedural uncertainty. Couldn’t you ask it to respond y/n in response to a question and look at the next token sampling probabilities?

Unsolicited endorsement, but I bought a “Brick” device recently, and I can’t recommend it enough for cutting back on social media. Infinite scroll really destroys your ability to concentrate.

Would love some kind of randomized A-B test where CEOs are replaced with AI and the outcome measured. Would not at all be surprised if it’s far better than the bottom quartile or more.

More practically, I think you could give a board of directors access to an LLM that sees all levels of company operations and let them talk to that chatbot as a stand-in for the CEO on a 24/7 basis.

Immediate page abandonment for me. I would appreciate more pledges of “this is entirely human writing, my own thoughts, imperfect as they may be”.

Using a better tech stack also provides opportunities to simplify.

Man, I hate that tech has just become money. Content like this perpetuates it.

I published in Nature Physics and the copy-editing process was quite embarrassing, to the point where we had to repeatedly nag them to stop them from making the manuscript presentation worse.

To be clear, I’m not talking about subjective style issues, I mean conforming to their own spec and avoiding careless bugs.

All remaining work fell on the backs of the physics referees. I’m not sure what value Springer provided from an editorial standpoint. It was disappointing to say the least after all that hard work.

Agree on article frustrations. Perhaps a better explanation, skills are just disk-cached prompts conditioned on verified success. The conditioned on verified success part might seem inconsequential, but it’s the whole thing that gives skills their value. Also the fact that their loading can be scoped to a certain calling context.

If anyone from Kalshi or Polymarket is reading this: your advertising is reprehensible. Dumb bros bragging about getting rich betting on the weather. Cringey get-rich-quick influencers. No one believes for a second your claims that prediction markets serve the public good. I’ve never wanted an industry to fail so badly.

It’s my opinion that societal rules should be derived from more fundamental virtues and notions of morality. AI is a capability, and it can be used in moral or immoral ways, but it’s more like a knife than an assault rifle. I don’t want AI forced down my throat by SF bro evangelists, but I also don’t want to see it banned as a useful technology. I wish people didn’t feel the need to adopt extreme positions on this topic and were capable of advancing more nuanced perspectives.

This, creating order from chaos (reducing entropy) is difficult and requires real intelligence. Inflating some small prompt into a wall of text and creating a bunch of entropy in the process is not as useful as it appears.

Yes, we can crawl our entire internal documentation via LLM. Want to know if someone is already working in the space of your latest idea? Ask Claude, it hits the internal search APIs and finds docs and references directly relevant to your query. There are a lot of separate document stores so this took a lot of effort previously. I can also query Slack, Outlook, etc. I don’t understand the cynicism in your comment.

I have mixed thoughts on this. These thoughts are my own. On the one hand, it’s objectively silly to pretend like we’ve solved the age old problem of measuring developer productivity. Metric-obsessed leadership can also be intolerable, counterproductive, and it’s a good way to paint yourself into a corner undervaluing your best talent and overvaluing your mediocre talent.

That said, I’m kind of having a blast using CC in corporate with all the connectors available at our disposal, and I baffled how little some of my coworkers know about what’s available and what the capabilities are. So it’s clear that perhaps some encouragement is prudent for those who are slower to embrace new technologies, but I’m not sure tokencounting and tokenmaxing are the answer.

There’s nothing wrong with that and perhaps this isn’t useful for you. Reflecting on my own socialization, some of it is spontaneous and some of it is “it’s been a while since I’ve seen so and so, I should reach out”. This app is more for the latter.

It’s also a good ledger of exactly how much I socialize and with whom. Often I find that I think I hangout with certain friends more than I do, and this helps me confront the reality that I need to put in more effort.

To be clear the catchup frequency is not meant to promote “hangout every X days” but rather to encourage me to be more mindful with who I hangout with and how often (roughly).

I understand your sentiment, and there’s absolutely some truth to it, but I don’t think the path forward is throwing more management resources (or layers) at the problem. And I don’t think the management technology industry is the answer either.

I think the solution will be small (1-5 person) teams where product and engineering sit next to each other and have clear authorization to launch directly to prod at their discretion. The gripes about performative work tracking mechanisms and the realization that non-tech considerations are now the bottle neck are not mutually incompatible.

Loved my Filco stealth, thanks for the great products!

I physically twitch every time I hear a flywheel mentioned. Intended to be evocative of certain physics without actually substantiating any of it.

I’m not fully subscribed to the idea that all public scraping of websites is bad, or that this project is a productive contribution. It would be nice to have search engines other than Google, and that necessitates bots being able to index your website (in some respectful manner).

It’s depressing how many lucrative big tech (FAANG/Unicorn) jobs are effectively scams with business glitter on top. And it’s not always obvious unless you really sit down during the interview process and scrutinize the business model.

You took the words out of my mouth. In a megacorp, AI multiplies into about 10% of my work and 10x’s it making me roughly 10% more efficient. When I use AI for side projects and don’t have to work with a bunch of stakeholders, dependency owners, and opinionated management, that 10x multiplies into my full effort and the project moves 10x faster.

The intersection of investors who understand what it takes to build a game engine and also the capabilities and plausible future capabilities of Google’s model is practically zero.

The tail of the distribution justifies the entire distribution. I agree that a large percentage of PhD research is inconsequential, but a small percentage is massively consequential. It’s ok to whiff on a thousand STEM PhDs if you pick up one Andrej Karpathy (for example).

There’s more to AI than foundation models. I think you are going to see meaningful progress on chore automation over the next decade through a combination of algorithmic and mechanical improvements, and it will measurably improve our lives. Recently got a Matic robot (awesome btw), and I no longer feel the need to vacuum my floors. It’s not life changing, but it’s an appreciable convenience upgrade. The capabilities feel like a peek into the future.

The customer service is better. Rates are generally way better. The downside is generally their phone apps and their fraud departments. I had a lovely credit union but their CC was basically unusable because they hired some crummy and over aggressive fraud detection third party that would lock down my card any time there was a strong breeze.

Their tool code use makes a lot of sense, but I don’t really get their tool search approach.

We originally had RAG as a form of search to discover potentially relevant information for the context. Then with MCP we moved away from that and instead dumped all the tool descriptions into the context and let the LLM decide, and it turned out this was way better and more accurate.

Now it seems like the basic MCP approach leads to the LLM context running out of memory due to being flooded with too many tool descriptions. And so now we are back to calling search (not RAG but something else) to determine what’s potentially relevant.

Seems like we traded scalability for accuracy, then accuracy for scalability… but I guess maybe we’ve come out on top because whatever they are using for tool search is better than RAG?

Less confusing title would be 83% of those killed are estimated to be civilians.