chrismustcode

If your USB ports are disabled for data how do you plug in a mouse and keyboard?

Do you mean kwh? Your per Watt is way off.

Also France is around 7c (euro) per kwh from nuclear generation. If the whole west was all in on nuclear for last 50 years it would have resulted in lower per kwh costs.

You can get 6.1 sol on ultrafast.

For some reason though on openrouter the tps of 6.1 sol ultrafast is usually lower than opus 5.5 fast. Is there readings off as surely cerebras hardware is faster than low user density nvidia still?

https://openrouter.ai/openai/gpt-6.1-sol

Lambda Web Functions, per the article, and Vercel Fluid Compute also do this.

https://vercel.com/fluid

MicroVM which I am presuming are firecracker

It's not.

Google broke ground this year on a owned and operated DC near Stockholm but GCP Stockholm right now is third party DC with their own racks etc.

Curious to see if they change that when it comes on line, or if the just add an AZ.

If Warsaw is just a cloud region it may not be on this site.

This is a Google owned and operated data centres.

Some gcp cloud regions are just rented floorspace in DCs from equinox etc.

They won't show on here if the latter.

A single AWS region can be a dozen DCs. An AZ is shared infra (power circuit, fiber ingress) and be over a few buildings.

La marzocco micro and mini are still high effort to make your coffee compared to some automated 90% cheaper delongi.

I think people sway to low effort endeavours that still have a reward at the end (even if it lesser reward than high effort).

You might as well use Opus for everything there.

Changing model would be cache busting spiking usage for no good reason when Opus can do it all.

Haiku 5.5 might fit well though depending on pricing.

Cache read is the same as Opus as well where most agentic workflow cost comes from.

Not quite sure where this fits well. Maybe small one one off requests like using Claude desktop/web?

It's soul destroying going from personal computer Linux (been figuring out the distro for me) to Windows for work.

Even worse they use Office 365 with sharepoint, they've even turned off web search somehow on copilot and can only search internal servicenow and SharePoint pages making it worse.

Things are pretty dire.

Anybody hiring who uses Google workspace and (preferably) Linux or MacOS let me know.

TBF they did have round ones in the premier league at one point - https://www.reddit.com/r/PremierLeague/comments/tpjqcr/who_e...

Apart from SpaceXAI no?

There's still a massive compute crunch, I know the opencode guys had been struggling to get capacity, Claude effectively lowered it's limit till the SpaceX deal, Google is struggling.

https://x.com/thdxr/status/2024539643673211054

https://www.anthropic.com/news/higher-limits-spacex

https://finance.yahoo.com/technology/ai/articles/ai-demand-o...

You don't consider Input $0.435 Output $0.87 cache read $0.003625 per million tokens for near frontier intelligence cheap?

I once got told for an internal promotion I couldn't put anything regarding my current role, responsibilities and achievements in the role. I got told to put any volunteering or previous.

Reason given was it's what is expected at work everything you do in your role, you need to show above and beyond.

If they can block IPs of cloudflare what extra mechanisms would be needed to block VPN IPs?

Is aider even a thing considered anymore?

It was pretty much first for CLI agents and had a benchmark that was the go to at the start of LLM coding. Now the benchmark doesn't get updated and aider never gets a mention in talking about CLI tools till now.

5 days ago OpenAI raised $122b and 26' Q1 recorded the largest amount of startup funding in a Q ever.

I wouldn't say it's drying.

https://x.com/OpenAI/status/2039085161971896807

https://techcrunch.com/2026/04/01/startup-funding-shatters-a...

I'm on sky in the UK which is marked as not safe due to no RPKI.

It's not on the list so imagine there is a fair few missing, would be neat to have a table you could filter by country, provider type (cloud/isp etc) based on real results from users.

edit: there's a show all button to expand the table

This isn’t accurate even for API prices for a request/response.

Go on something like openrouter with gpt 5.1 and use the chat then check the billing and you’ll see an average joe query is like 0.00102 or something.

You’re quoting figures from articles for initial ChatGPT release in 2022

They were failing as an online IDE for several years then growth shot up after the AI pivot.

Couldn’t they just send some hardware down Texas to co-locate there (presuming specialist hardware) and add another deployment target for their software? Would it be that hard?

I thought they use GPU for learning and TPU for inference, I’m open to been corrected.

I’ve never had an experience in any house or office where there’s been enough sockets to leave everything plugged.

I honestly find in go it’s easier and less code to just write whatever feature you’re trying to implement than use a package a lot of the time.

Compared to typescript where it’s a package + code to use said package which always was more loc than anything comparative I have done in golang.

A lot of these have business logic literally in the database built up over years.

It’s a mammoth task for them to migrate

I agree it should be across the spectrum where people have the same rights to privacy.

> those who can afford to make the government look the other way and those that can't.

> Congratulations to Apple on lobbying for its own money. Very noble.

But what’s your implication here, that Apple shouldn’t have fought it?

You hit the limits on this like immediately (it’s 1000 requests which each tool call is a request and even then I’m convinced pro isn’t 1000).

Also data retention Gemini has which you have to download a vs code extension to turn off for the CLI.

It isn’t an enterprise product it’s a way to get data for tool calling for training as far as I see it (as it currently stands).

I’d be stunned if a 270m model could code with any proficiency.

If you have an iPhone with the semi-annoying autocomplete that’s a 34m transformer.

Can’t imagine a model (even if it’s a good team behind it) to do coding with 8x the parameters of a next 3/4 word autocomplete.