andrewcamel

Re per token, I had the same reaction, but given both labs are economically advantaged moving customers to per-token consumption... almost want to avoid this on principle. Even if not intentional, benefitting from a degraded product is not something I want to accept or enable.

More now than ever (since original ChatGPT release), the OSS models and open harnesses (eg Pi) are looking mighty attractive.

I’m not sure it will cause a reversal, but I bet the k8s takeover wouldn’t have happened with costs of today. You inherently end up using far more resources than you need - between HA, built in services, overhead per node, etc.

I personally will be using more resource efficient approaches in everything I do. Question is just what provides the closest set of benefits without the full k8s weight.

I’d imagine you could seed it more easily by focusing on top 50 routes by passenger count in domestic USA. Then go from flight schedules for top airlines into tail numbers into body types etc.

Big fan. One feature idea/request - a map showing coverage with 0-100% by route (red/yellow/green lines). I’m just curious to see where I should think to look for / expect starlink options. Probing into a few upcoming trips showed basically no coverage.

Philosophical question, but after reaching critical mass, should languages even aspire to more? I.e. do you risk becoming "master of none"? What's wrong with specialist languages? I.e. best of breed vs best of suite?

I agree with author Go is getting squeezed, but it has its use cases. "COBOL of could native" implies it's not selected for new things, but I reach for it frequently (Go > Java for "enterprise software" backends, Go > others for CLI tools, obviously cloud native / terraform / CI ecosystem, etc.).

However in "best of suite" world, ecosystem interop matters. C <> Go is a pain point. As is WASM <> Go. Both make me reach for Rust.

What response would you suggest for Israel?

My understanding is their civilian vs combatant ratio is lower than comparable urban conflicts. So it’s hard to respond “softer”.

Population adjusted, 10/7 was 10x+ worse than 9/11. Challenging to expect zero response.

So what else is there?

Are tariffs already in place or is this just a thinly-veiled scapegoat for haircutting traffic allocation by 95%? To a customer, it certainly feels like a bait and switch to sell a subscription product and once customers are embedded materially change the economic trade.

It's translation (text -> text), not speech -> text.

Do they report the % of revenue from AI / data center chips vs all others? Or is there a way to derive it? Surprising to hear the slump given what we're seeing in AI-driven GPU demand. I guess that's just not as meaningful a portion of the TSMC business as I would've thought.

It's a collection of interviews posted in the form of a library. So bit blog-like in structure, but just a collection of ideas on how to leverage this new tech.

I'm starting to outline them here: ctlresearch.com . Upcoming interviews with Chief Architect at Intuit, Head of Procurement at DoD, etc. DoD already shortened process of writing structured "requests from industry" from 3 months to 1 day. Makes it far easier to get requests out to vendors. Next step is an auto-complete bot that helps vendors respond with required language to RFPs.

I have 20 interviews coming down the pipe -- all of which have highly tactical / near term valuable ideas like this.

Maybe not the right term... Just that a lot of other libs act like guardrails, i.e. let the model generate what it does (in full form text / GPT output), and try to then parse out what you want, error if output doesn't conform to standard format. As opposed to basically only allowing the model to generate into the already-built JSON form fields. Understandable why this guardrails/parsing approach is so popular though... can't do what this library is doing with OpenAI API. Need to be able to manipulate the token generation; otherwise you're forced to take full text output and try to parse it.

Seen a lot of things trying to do this by pressure testing the outputs, but all feel like anti-patterns. This is the first that seems like the "right" way to do it. Better to manage how the model is generating vs creating one more potentially faulty "glue" layer.