Hacker Newsnew | past | comments | ask | show | jobs | submit | bravura's commentslogin

I know this is a general question, but:

Assuming a US startup is considering engineering hires outside the United States, how does one currently assess the likelihood of getting them a visa to work in the USA? And what timeline and cost would be involved?


Unfortunately, a case by case analysis would be required. However, if they are from a country with its own visa (that is, Australia, Canada, Chile, Mexico, and Chile), it's relatively easy to get engineers visas.

Agreed, to me they are very different takes on what is a punk attitude.

Palahniuk: Underneath the veneer of the banal, you will discover everything is rotten and sycophantic but somehow tender and relatable.

Bourdain: Underneath the veneer of the banal, you will discover an honest struggle for something far more respectable than what is typically venerated. Eat their food, dance to their music, and you will enjoy.


Isn't the challenge that introspecting graphql will lead to either a) a very long set of definitions consuming many tokens or b) many calls to drill into the introspection?

In my experience, this was the limitation we ran into with this approach. If you have a large API this will blow up your context.

I have had the best luck with hand-crafted tools that pre-digest your API so you don't have to waste tokens or deal with context rot bugs.


Well either that or stuff the tool usage examples into the prompt for every single request. If you have only 2-3 tools GraphQL is certainly not necessary - but it wont blow up the context either. If you have 50+ tools, I don't see any other way to be honest, unless you create your own tool discovery solution - which is what GraphQL does really well with the caveat that whatever you decide to do is certainly not natural to these LLMs.

Keep in mind that all LLMs are trained on many GraphQL examples because the technology has been in existence since 2015. While anything custom might just work it is certainly not part of the model training set unless you fine-tune.

So yes, if I need to decide on formats I will go for GraphQL, SQL and Markdown.


Call your father. And just chit chat for ten minutes.

Thoughts on any-llm-gateway versus litellm-proxy?

litellm is a great library, but one team using litellm-proxy reported having many issues with it to me. I haven't tried it yet.


What were the problems? I've been trying it out and haven't hit issues yet, but not using it at scale yet so I'm curious what to watch out for. I figure it's open source (MIT) so I can make changes as needed if there was anything particulary annoying.


Yeah, I wonder what gaps in Litellm Proxy made Mozilla want to even do this.


LLMs are not deterministic though. So by definition MCPs are not deterministic.

For example, GPT-5 doesn't support temperature parameter. And even models that do support temperature are not deterministic with temperature=0.


How large is a lion?

Learning the size of objects using pure text analysis requires significant gymnastics.

Vision demonstrates physical size more easily.

Multimodal learning is important. Full stop.

Purely textual learning is not sample efficient for world modeling and the optimization can get stuck in local optima that are easily escaped through multimodal evidence.

("How large are lions? inducing distributions over quantitative attributes", Elazar et al 2019)


> How large is a lion?

Ask a blind person that question - they can answer it.

Too many people think you need to "see" as in human sight to understand things like this. You obviously don't. The massive training data these models ingest is more than sufficient to answer this question - and not just by looking up "dimensions of a lion" in the high-dimensional space.

The patterns in that space are what generates the concept of what a lion is. You don't need to physically see a lion to know those things.


> How large is a lion?

Twice of half of its size.


Can you be more specific about "size" here? (Do not tell me the definition of size though).

You are not wrong though, just very incomplete.

Your response is a food for thought, IMO.


Have you tried e2b or Daytona fast start vms?


How do you use a non Anthropic model with Claude Code?


they have a Claude Code specific endpoint...see the excellent docs https://docs.z.ai/devpack/tool/claude


Why don't companies have an about page on their home page any more? How an I supposed to know what z.foo is really about? Random page vibes?

</rant>



I use liteLLM local running on Docker.

https://www.litellm.ai/


The text-based software that would eat work management is one that embraces the incumbents rather than avoid them.

I want a bidirectional SaaS <=> YAML/JSON adapter. So that I can push and pull our CRM (and other SaaS utilities like project management) into a common (schematized) YAML format.

The YAML then can be analyzed and modified using LLMs and/or stored in git.

And then use the bidirectional sync to reconcile conflicts and push.

So I can do work processes on the console, and still collaborate with people who want the native web UI.


Agreed, this is on my mind as well.

Thinking of Terraform, you have data blocks that can grab data from an external source. Still trying to grok what would be a convenient way of doing something like this - whether that gets generated to DSL, or if data pulled in dynamically as you build the org graph...

Having your plain-text workspace as a unified structural source where you pull in data from external systems would be potentially powerful.


You'll unfortunately have to agree on some sort of state representation for each source and then delve into those APIs to extract that information


Hi there, I've made my own text-based todo list with compatible web view which is kind of similar to what you did. (though far from completion)

There's a sync engine behind it so the UX is extremely responsive.

Link: https://mglogi.com/portfolio


I’d be happy with SQL access, which I think gets to roughly the same place.

I’ve done something like what you’re talking about before for a CMDB, though it was one way YAML -> DB sync. Many to many relationships were a pain to view, there’s not a great way to put them in YAML that makes them easy to read. Can’t embed them because then you have multiple copies and which one is the real one. References suck because you can’t see the relationship and the related objects at once.

The real killer is permissions, though. Your sync tool basically has to have admin privileges, which means permissions have to be checked at merge time, and then you’re rebuilding the entire permissions flow as a git hook.

SQL with RLS is capable of implementing permissions in a way that works for both API access and direct SQL access. I get the feeling few companies do it, but they could.


I would like the native web UI when I am being too lazy, and terminal when I am in the zone..


Huh. Could those be git submodules? Like you’d have all your personal state in foo/ and then clone the work state into foo/work/ .


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: