This particular example aside, I don’t think it being derivative and simplified is necessarily bad. Libraries that are popular today were written for humans and reinforced by LLMs via training. It’s unlikely they represent the ideal interaction surface for an agent.
There was a study recently that LLms prefer resumes written by LLMs rather than by humans. Stands to reason they would prefer apis written by LLMs.
This is probably the early days of such intentionally simplified agentic semantic primitives like “DAG Workflow” where the answer for why not Temporal is that LLMs prefer different things than humans.
Is the difficulty that in high entropy situations, you can’t really tell whether it’s because the model is uncertain, or because of the options are so semantically similar that it doesn’t matter which one you choose? Like pure synonyms.
I’ve been experimenting with LiveStoreJS which uses a custom SQLite WASM binary for event sync, so for simplicity I’ve also used it for regular application data in browser and found no issues (yet). It surprised me that using a full database engine in memory could perform well vs native JS objects at scale but perhaps at scale is when it starts to shine. Just be wary of size limits beyond 16-20mb.
Of all the interface modalities available, CLIs seem like the most natural for copilots to work with. Lots of examples in the training data, universal interface for help, maps well to the sequential nature of token generation, similar syntax for different OSs… I can see them replacing skills and MCP et al from the model’s perspective.
How do they prove this? It sounds like the plaintiffs basically claimed they were rejected a bunch of times and since the resume had recognizable indicators of protected classes they must have been discriminated against?
Don’t get me wrong, I do this work, and Workdays statement of “we don’t use protected classes” instead of “we test our models to prove they are unbiased when given recognizable indicators of protected classes” is pretty telling. Because it’s hard and if you solved it you would be proud. If you don’t control for it it WILL discriminate. See Amazon’s experiment a decade ago.
I’m just really curious how all this plays out in front of a judge.
Different experiences leading to an inability to communicate/empathize is universal, I think:
> 326. “I can’t know what is going on in him” is, above all, a picture. It is the convincing expression of a conviction. It does not give the reasons for the conviction. They are not obvious.
> 327. If a lion could talk, we wouldn’t be able to understand it."
1. The first computers were huge machines that everyone logged into with thin clients
2. Then the PC era where computing became decentralized and all computing took place on a single machine
3. Then the internet, which went back to the thin clients connecting to large central servers
If you believe that the reason for the shift away from decentralization was to enable coordination of multiple PCs so they had access to the same centralized data (for reading or shopping or whatever) then its really not a stretch that distributed consensus algorithms could be a way back to decentralization.
> and no, blockchains don't count for now - if anything, cryptocurrencies are a lesson in just how quickly systems centralize
Not disagreeing with you on the current state, but I believe that is because the people most engaged right now are the same types who day trade penny stocks and options, i.e., low-information high risk seekers. Many of the structures of the cryptocurrency market seem lifted from a student's textbook about how markets can or should work. I see zero discussion of why different structures like open outcry exchanges and centralized regulatory agencies existed originally and what roles they serve now. Maturity, both in the users and the infrastructure, should help this.
I recently had a discussion with a trader-type about why exchanges should exist at all for cryptocurrencies given the volatility cause by a combination of high volume, few market makers, and lots of friction between exchanges. The responses were about price discovery and how would we trade? I don't think they believed me that, as a stock trader, you have almost certainly never traded a stock on an actual exchange and high volatility means its value for economic price discovery is nearly useless.
But please resist the urge to conclude that what is happening in the cryptocurrency markets has any bearing on the potential of distributed consensus algorithms.
Youtube? If you ever watch a kid going through one of those toy unwrapping channels its not far-fetched to think that there's something going on there that's playing on the way their brain works to keep them coming back for more.
You may find this article interesting that criticizes the abstinence-based approach of Alcoholics Anonymous. Research has shown that cognitive behavioral therapy, motivational interviewing and medication to control withdrawal symptoms is more effective for a large population than absolute prohibition. In fact, I wonder if there are medications that address withdrawal symptoms from social media?
There was a study recently that LLms prefer resumes written by LLMs rather than by humans. Stands to reason they would prefer apis written by LLMs.
This is probably the early days of such intentionally simplified agentic semantic primitives like “DAG Workflow” where the answer for why not Temporal is that LLMs prefer different things than humans.