smcleod

16 ms·

331.
by smcleod 2y ago
Hey, is the source open for this?
332.
by smcleod 2y ago
I have never used an LLM for fictional writing, but I have been writing large amounts of code with them for years. What I'd recommend is when you're defining your plan up front as to the sections of the content, simply state in wh
333.
by smcleod 2y ago
Unless I'm misunderstanding this - Cline can do this off the shelf (no need to install & subscribe to cursor), I'm surprised you can't do this in cursor actually?
334.
by smcleod 2y ago
Siri's voice recognition has gone down a lot over the last 5 years actually, it used to be far more accurate, its speech has certainly improved to be more natural sounding. Its capabilities though... yikes
335.
by smcleod 2y ago
There's so many MCP clients out there, Cline is IMO the best agentic coding tool that makes heavy use of it.
336.
by smcleod 2y ago
No mention of how well it's claimed to perform with tool calling? The Gemma series of models has historically been pretty poor when it comes to coding and tool calling - two things that are very important to agentic systems, so it will
337.
by smcleod 2y ago
It's great! Easy to work with, makes it quick to build tools and isnt over complicated.
338.
by smcleod 2y ago
Oh, sorry what do they/you call them? (The storage above the rear axil at the rear for the car)
339.
by smcleod 2y ago
LiteLLM requires you to call them to discuss pricing to enable SSO... sure part of it is open source but not the part you need to scale it.
340.
by smcleod 2y ago
16GB? In 2025 that seems very small for anything other than gaming.
341.
by smcleod 2y ago
GPT 4.5 is insanely over price, it makes Anthropic look affordable!
342.
by smcleod 2y ago
Especially not one owned by Bezos
343.
by smcleod 2y ago
IMO the only good feature in it is the notification summaries, they're pretty accurate and useful. Siri on the other hand is embarrassingly bad.
344.
by smcleod 2y ago
It should be noted here they're talking about cold water (above freezing) - not ice baths which is what a lot of people seem to think of and is quite a fad in the "sports medicine" industry.
345.
by smcleod 2y ago
Actual is cool but its lack of solid integration with the open banking standard (as is used in Australia and other countries) made it a bit annoying as its partnered integrations are pretty rough.
346.
by smcleod 2y ago
Just 4K. Because deepseek doesn't allow for the use of flash attention it means you can't run quantised qkv
347.
by smcleod 2y ago
I get around 4-5t/s with the unsloth 1.58bit quant on my home server that has 2x3090 and 192GB of DDR5 Ryzen 9, usable but slow.
348.
by smcleod 2y ago
That qkv PR was mine! Small world.
349.
by smcleod 2y ago
Stats is great! The only issue I've had with it is each time it updates itself as it's unsigned the binary is flagged and can't be opened until you run xattr -rc /Applications/Stats.app on it.
350.
by smcleod 2y ago
Nice work! Love that it's a native app and not another Electron Chrome wrapper. Well done!
351.
by smcleod 2y ago
The responses from the DeepSeek hosted API are far more censored that if you run the model yourself.
352.
by smcleod 2y ago
I think most of the model creators share their model usage examples so high at 0.6-0.7 simply because it's what a lot of the client apps use. IMO this is WAY too high unless you're doing creative writing. Generally I set temp to 0
353.
by smcleod 2y ago
Yep that's pretty much what I did, their calculation for the layers was slightly off though, I found I could offload an extra 1-2 layers to the GPUs
354.
by smcleod 2y ago
min_p is great, do you apply a small amount of temperate as well?
355.
by smcleod 2y ago
I just ran it up on 48gb (2x 3090) + overflow into CPU RAM and it runs at around 4tk/s (only a little 8k context size though) which while absolutely not something I'd personally use daily - it is actually usable.
356.
by smcleod 2y ago
While censorship and political bias is of course bad, for a lot of their intended use cases you're really not going to hit up against it. Especially for text to image and coding models (deepseek, Qwen and other Chinese models main stre
357.
by smcleod 2y ago
Can vouch for cotypist - it's great!
358.
by smcleod 2y ago
Congrats on the launch, QwQ is fantastic!
359.
by smcleod 2y ago
FYI - website seems broken, docs link is invalid, readme mentions docker-compose but the compose file is missing, pnpm run build:server is broken, pnpm run start:app launches the app but the login fails as it seems it needs a database serve
360.
by smcleod 2y ago
Commute, noise, lighting, desk setup, open plan offices, hot dealing and generally having to sit in some corporate office rather than my own home. I'm an adult - I'll work where I want and that will never been an office ever again

More ›