Simi

digital hunter gatherer Thinking driven development

[Simi @sudosimi]

@ATaylorFPGA Can’t see the input for entering email on the page. Scrolls down but no input

[Simi @sudosimi]

@Jason______A 10 year old luggage that I can replace the silicone wheels. They offered “lifetime warranty” and when I went to the shop they said it’s not produced anymore and can’t help. Anything equivalent is 3 times the price. Zipper broke and persuaded a tailor to fix it.

[Simi @sudosimi]

@Pat_Stedman People think they have standards when those “standards” are unfounded demands.

[Simi @sudosimi]

@Ekaeoq Hanhart 417 39mm reverse panda

[Simi @sudosimi]

@2147mill With inflation I’m aware even the tax free ISA allowances are shrinking every year.

[Simi @sudosimi]

@ClementDelangue @huggingface Add TurboQuant and I’m running 196k context with similar VRAM card.

[Simi @sudosimi]

@PLeonard61032 @EC2man I made an outrageous claim so you can correct it

[Simi @sudosimi]

@DellAnnaLuca Most environmentally obscene thing I’ve seen in the US is irrigation in desert areas for golf courses.

[Simi @sudosimi]

@IlirAliu_ I recommend a linear algebra intro or refresher before going for this book. Exercises will be a pain without that.

[Simi @sudosimi]

@BowTiedAsset @FrenchOG3 And I’m sure women wait too for the one, instead of supplying what they need from the free market of dating apps. I can’t believe the solved the problem for everybody by selecting for men’s scarcity and desperation.

[Simi @sudosimi]

@CraigSpillard With this calculation the landlord looks like an intermediary for the state, not an independent economic agent

[Simi @sudosimi]

@GreatBritishTT The state probably makes 50p on every 1£ in at supermarkets.

[Simi @sudosimi]

@sudoingX subjectively the tradeoff for me is better than larger context with lower kv cache, keeps doing a good job after the first and second compaction, still not satisfactory quality with q4 kv cache and larger context. Most tasks I give it fit within 128k context.

[Simi @sudosimi]

@sudoingX decided for a similar setup on laptop 5090:60k context q4km, q8 cache and doing the compaction on a laptop 4090 with nemotron 4b at f16. I get slower speeds 20-30tok/s and large contexts take 5-9 min to be processed

[Simi @sudosimi]

@_Unknown_D_ It’s voting economy. It’s the amount their potential voter base would consider a “fair” amount.

[Simi @sudosimi]

@LilaRest @huggingface @outsource_ What context size?

[Simi @sudosimi]

@mr_james_c Another case of something brilliant a century ago, but let’s not do it anymore. This is not managed decline it’s managing the place into decline.

[Simi @sudosimi]

@sudoingX Laptop 5090 with 24G vram - close to 3090 in memory bandwidth.