A nice addition (unless I missed it) would be to add an existing API key for remote model access?

Interesting stuff. Unsure about the Phlex approach but could be worth a go.

Didn't have any luck with the Date Picker component demo page (no pop up or validation?), and that's my usual go-to for seeing how these libraries are doing.

Super interested in this when the eye tracking is available but need some reviews for that first. For sit-down VR the price isn't really an issue for this, given how the GPU market is now anyway. It is a niche product for sure.

Foveated rendering with decent eye tracking could help us get out of the combo of high resolution / low framerates for clarity vs low resolution / high framerates for comfort.

Also, Valve seem to be ready to say something but it feels like it'll probably be (sensibly) a Steam Deck strapped to your head for stand-alone solution, and that's not the same market as this PC VR set-up is aimed at. Maybe if Valve take the 'store subsidy' hit and price it well and you can run a cable to it anyway..

This bit feels like we are being pushed away from the existing API for non-technical reasons?

> When using Chat Completions, the model always retrieves information from the web before responding to your query. To use web_search_preview as a tool that models like gpt-4o and gpt-4o-mini invoke only when necessary, switch to using the Responses API.

Porting over to the new Responses API is non-trivial, and we already have history, RAG and other things an assistant needs already.

From their perspective, if they don’t have your data, it’s too easy to switch providers.

Exactly this is what's going on. Moat-building.

I can’t find that text in the announcement. In fact it sounds like you have to use a specific model with the chat completions endpoint to get web searches.

In the API they are named like "gpt-4o-search-preview".

It would have been nice if the Completions use of the internal web-search tool wasn't always mandatory and could be set to 'auto'. Would save a lot of reworking just to go the new Responses API format just for that use case.

Some rough cost numbers of using this would be nice.

With the exception of Cache Reserves, all these features are free.

Cache Reserve pricing is based on storage usage but optional.

Caching is free up until a service ceiling - certain file types and sizes are excluded in the free tier for example.

One side effect that I don't think I like of these tools is the inevitable push for developers to just use the most popular languages. Javascript and python are the main LLM trained languages, so it just self reenforces that it is best to use those for everything. There is something a bit sad about that. I guess the dream is that they are just intermediatory data anyway, like a chatty bytecode layer.

I tried this and couldn't get it to work on a Rail 7.1 Ruby 3.2, no output. The readme.md could do with some examples of what sort of recommendations come out. Edit: it outputs metric, just no recommendations so moving on.

Here in Canada, there was a trend for the 'Couch Potato Portfolio' with the arrival of popular ETFs that used these splits between equities and bonds and then just rebalancing once a year. The key point was really using entire index ETFs with low MER fees on auto, with the bonds portion as the stability. 'All in' ETFs have sort of replace that.

My first 'job' as a kid was working at Oasis Software on White Lightning, which was a Forth based game making kit for the ZX Spectrum (amongst other platforms). About a million years ago now. Or now ago million, if you know Forth. https://worldofspectrum.org/archive/software/utilities/white...

No, in Forth it would probably be:

    1 MILLION YEARS AGO NOW

In other words, push 1 to the stack, multiply TOS by 10^6, multiply TOS by 86400*365, negate TOS, and finally add TOS to the current time in seconds.

Forthers did this kind of thing all the time -- invent Ruby style DSLs based on Forth's stack semantics.

White Lighting was really a sprite editor packaged with an interrupt driven DSL for games, but back before DSL was a term. Forth was a good choice for a compiled target on the platforms back then, given the constraints. Ironically I do ruby now.

I've picked up Forth to code for Agon Light, which is a new spin on those old platforms - it's BBC Micro-patterned, but uses some modern chips to make things faster and roomier. Easy to develop for on emulator too. One of the complaints I've read about Forth on micros was that even 8Kb interpreter overhead would be too much for a commercial game - what I've gathered is that many people used it primarily as a macroassembler in that context. But I have 512Kb and 20MHz to play with, plus external display processing, so I can probably ignore that overhead and do just fine.

I find Forth quite a bit easier to work with when I treat it as an interactive stack machine and do the bulk of development through offline code generation. That gives me a best-of-both where I can do useful things interactively, but I can add compiler tech to help rein in and structure whatever I'm doing.

I frickin' LOVE the Agon Light. To me it's the most exciting and promising of the "retro-inspired new computer" designs.

I've messed about with Forth on the Commodore 64 (via THEC64). It's fun but I haven't dug too deeply into it, like building a game with it or anything.

I play around with White Lightening occasionally because of the ease of making games with it on C64 and the comprehensive documentation (PDFs of it have been preserved).

I had the machine lightning version for the Commodore 64. I remember that the manual was printed red paper so you couldn't photocopy it.

Yep, the tape to tape copying days were really happening around 1984 so the manual in either red or green paper was the only way to keep it selling. The sprite library package and Forth's closeness to machine code were its strengths. Not easy to work with though.

I didn't manage to build any Games with it, but I think it helped me to read Postscript files a few years later :)

I remember seeing adverts for this and wondering what it was. Was it used for any commercial games?

Some smaller ones, but nothing big. There was a user submission contest that got some traction, but the jump up from BASIC to using this was not easy. This was the time of 'rock star game devs' but interrupts, memory paging and sprite libraries probably crushed a few teenage dreams. An old video of what it could do https://www.youtube.com/watch?v=IgmZ2Ht-QiQ

From the manual (those were different times indeed!):

> We are also interested to hear of any extensions or routines you may develop, and if sufficient interest is shown we will start a News Letter, and possibly, even a User Group.

I remember the ads for it and wishing I could afford it. It seemed so awesome, but as you mention elsewhere I probably would've struggled.