fintechie

Well you know how it goes... you need to read between the lines. I can agree with you on your "principles", but not enforce them myself.

This kind of flows have been documented in the wild for some time now. They started to pop up in the Cursor forums 2+ years ago... eg: https://github.com/johnpeterman72/CursorRIPER

Personally I have been using a similar flow for almost 3 years now, tailored for my needs. Everybody who uses AI for coding eventually gravitates towards a similar pattern because it works quite well (for all IDEs, CLIs, TUIs)

This... OpenClaw is the best thing to happen to security and forensic firms since Windows XP. The amount of hacks, data/credential leaks, etc to come out of this will be of unfathomable proportions.

Aren't SOTA models doing this for few cents already? at least when I tell Claude to add a systems draw or arch flow to a README.md he'll do it quickly.

Can you? can you provide a historical (25+ years) chart of reservoir levels in Cyprus or any EU country? Otherwise let me assume you just fell for a sensationalist article

Italy, Spain... eg. https://embalses.net shows maximum historical levels after a sharp bump these past 2 years... pretty sure it's the same story in many EU countries. Droughts are there until they aren't. Normal fluctuations if you check an actual chart going 50-100 years back.

Can you post a historical chart of Cyprus? maybe it tells a different story

It's a crisis measured in hard numbers: reservoir levels, rainfall data, aquifer depletion rates.

Of course I went to check the actual numbers from official sources and they tell a different story. Reservoir levels near historical maximum. So much for building an article on "hard numbers" without pointing to sources.

It's not that it isn't better, it's actually worse. Seems like the big guys are stuck on a race to overfit for benchmarks, and this is becoming very noticeable.

Is this the first major flop from Anthropic? This thing is unusable. Slow, awful responses. Since Sonnet 3.5 the only real advance in LLM coding has been Gemini 2.5 Pro's context length. Both complement each other quite well so I'll stick to switch between these 2.

https://github.com/TabbyML/tabby

Quite good, it would sound much better with SOTA voices though:

https://github.com/nari-labs/dia

I've been using this one with Cursor the past few months...

https://github.com/foges/whisper-dictation

Yeah that may be true. But one can only ask who wins from all the polarisation. Co-workers at odds, friends at odds, even family members at odds. Tech is(was?) intended to be our "happy place".

IMO this is a bigger blow to the AI big boys than Deepseek's release. This is massive for local inference. Exciting times ahead for open source AI.

I don’t understand how stupid some traders can be.

90% of traders lose money, so that's a data point...

You're trying to apply rational thinking but that's not how markets work. In the end valuations are more about narratives in the collective mind than technological merit.

The parallels with the dotcom bubble are clear... Tech is magical. Tech becomes cheap/free. Stock holders panic.

For some time it's been clear that there's an AI bubble and this may be what finally pops it.

Gave it a go for several projects, but didn't like it... for big projects it gets messy, fast. It also feels like it has become the new bootstrap.

I'm very happy with my current CSS-in-JS workflow. Crafting good old css with LLM help. You just show the LLM a pic, ask for the components.... boom, done (with proper naming, etc)

Hopefully this makes the Cursor team reconsider security (which doesn't seem very good really).

Stopped using it for serious stuff after I noticed their LLMs grabs your whole .env files and sends them to their server... even after you add them to their .cursorignore file. Bizarre stuff.

Now imagine a bad actor exploiting this... recipe for disaster.

Hype died down along with "microservices" where queues made more sense.

Yeah, it shouldn't be too difficult to build this with python. I wonder why none of the popular routers like https://github.com/BerriAI/litellm have this feature.

Problem is running so many LLMs in parallel means you need quite a bunch of resources.

Top of line MacBooks or Minis should be able to run several 7B or even 13B models without major issues. Models are also getting smaller and better. That's why we're close =)

We're nearing a point where we'll just need a prompt router in front of several specialised models (code, chat, math, sql, health, etc)... and we'll have a local Mixture of Experts kind of thing.

  1. Send request to router running a generic model.
  2. Prompt/question is deconstructed, classified, and proxied to expert(s) xyz.
  3. Responses come back and are assembled by generic model.

This story now looks like a minor incident if you compare it to FTX.

You can currently do this in an M2 Max with ollama and a Nextjs UI [0] running in a docker container. Any devices in the network can use the UI... and I guess if you want a LAN API you just need to run another container with with OAI compatible API that can query ollama.. eg [1]

[0]https://github.com/ivanfioravanti/chatbot-ollama

[1]https://github.com/BerriAI/litellm

508

although it does require you to wear headphones so the bot doesn't hear itself and get interrupted.

Maybe you can use some sort of speaker identification to sort this out?

https://github.com/openai/whisper/discussions/264

Hungary and Slovakia, the only European countries that haven't allowed illegal immigration in, are at the bottom of that chart, while Sweden is getting rekt...

Legal immigration is of course fine and necessary, as long as the state and socials systems can assimilate it... (Japan is a good example). But articles like this show concerning data points in western countries with uncontrolled migratory policies.

very odd... this looks like some kind of forced takeover

Another potential winner may emerge soon:

"OpenAI and Jony Ive in talks to raise $1bn from SoftBank for AI device venture"

https://www.ft.com/content/4c64ffc1-f57b-4e22-a4a5-f9f90a741...

It's down again... https://status.openai.com/

This comment encapsulates perfectly what's wrong about the ts talibans... Astonishing to see how their minds can't comprehend that businesses can and do thrive with js.