Electricity (on continental US) is pretty cheap assuming you already have the hardware:

Running at a full load of 1000W for every second of the year, for a model that produces 100 tps at 16 cents per kWh, is $1200 USD.

The same amount of tokens would cost at least $3,150 USD on current Claude Haiku 3.5 pricing.

This 35B-A3B model is 4-5x cheaper than Haiku though, suggesting it would still be cheaper to outsource inference to the cloud vs running locally in your example

If you need the heating then it is basically free.

Only if you use resistive electric heating, which is usually the most expensive heating available.

When they bumped the context size up to 1m tokens they made it much easier to blow through session limits quickly unless you manually compact or keep sessions short.

This generation's internet worm?

> Tom pulled up the tool’s specification on his diagnostic display. This was always the first step: read the spec, not the code. Clearly this writer has never felt the frustration of CC telling them a feature was never a part of the plan, because it overwrote the plan and then compacted.

This approach doesn't give access from the hypervisor to your private keys it gives access to other tenants to your private keys.

Back in 1960 us early detection systems mistook the moon for a massive nuclear first strike with 99.9% certainty. With a fully autonomous system the world would have burned.

Not a big exactly but if pip doesn’t work it goes straight to pip —break-system instead of realizing it needs a venv

Also if a prisma migration fails it will say “this is dev it’s ok to erase the database” before rerunning the command with —accept-data-loss

Given how many projects I've seen that run prisma migrations on prod from their local CLI instead of CI...

This scares me.

Give Cluade Code a go. It still makes a lot stupid mistakes, but its a vastly different experience from pasting back and forth with chat gpt.

There's no free trial or anything?

You can play with the model for free in chat... but if $20 for a coding agent isn't effectively free for use case it might not be the right tool for you.

ETA: I've probably gotten 10k worth of junior dev time out of it this month.

The chat is limited and doesn't let you use the latest model. if that's representative of the answers I would get by paying, it doesn't seem worth it.

Im not crazy about signing up for a subscription service, it depends on you remembering to cancel and not have a headache when you do cancel.

It is ridiculously easy to create an album with Suno and push it Spotify. I'm surprised its only 66% TBH

Anna's archive has a great analysis of the Spotify data.

They identify a huge surge in tracks that few listen to after gen AI started.

The analysis is worth reading. The distribution is (Pareto)^3 ~99% of the tracks played are 1% of the catalogue.

1. Generate slop music nobody will ever listen to 2. ???? 3. Profit

It's actually:

1. Generate slop music no _human_ will ever listen to

2. Use a botnet to "play" this music en masse

3. Profit

This is a whole arms race, with companies (such as Beatdapp) specializing in detecting fraudulent plays.

Source: I work for a niche music retailer that struggles with the same issues on a smaller scale.