Don't worry we're gonna need all the TPUs we could get our hands on for training those specialized models
@theahmadosman
-
Closed source AI risks: no control, sabotage, manipulation
By
–
Gentle reminder that, in closed source AI from companies like Anthropic and OpenAI You have zero control over how the models behave, and they can – Quantize it
– Distill it
– Sabotage your work and data
– Hot-swap to a cheaper/weaker checkpoint
– Make the model manipulative
– -
Continual learning ensures open-source AI victory over Anthropic
By
–
Continual learning, among other things, will ensure Opensource AI wins Anthropic has no moat and that’s why they’re trying so hard to make their models so useless in allowing others to break that wall But they just don’t know that they’ve already lost
-
Open Source AI Must Win as Civilizational Infrastructure
By
–
Opensource AI Must Win AI is a civilizational infrastructure for work, education, science, software, creativity, public services, and national capacity Access must not depend on closed APIs, remote platforms, shifting terms, opaque moderation, model availability, or prices set
-
Allegation that Anthropic nerfed models and Dario sabotaged codebase
By
–
Imagine how long Anthropic have had their models nerfed on purpose Dario masterclass in sabotaging your codebase while you were thinking Claude Code is just working for you lol
-
Loss of trust in AI benchmarks due to training on desired outputs
By
–
How could anyone trust this guy, his work, benchmarks, or any kind of output / product in Local AI after this is beyond me at this point “Any good benchmark will be trained on since that’s what people want.” I rest my case
-
Gratitude for focus on real problems over wrappers and agents
By
–
I am glad more and more people are starting to tackle the actual real problems rather than building wrappers and “agents”
-
Learning Path for LLM Serving Engines: vLLM, SGLang, TensorRT-LLM
By
–
How to go about learning all of this? 1st: Start with the serving engine view – vLLM: PagedAttention, continuous batching, prefix caching, CUDA graphs – SGLang: RadixAttention/prefix reuse, speculative decoding, MoE, structured/agent workloads – TensorRT-LLM: NVIDIA peak
-
User dismisses local open source critics as paid or stupid
By
–
I keep getting mentions on certain posts talking about how local / opensource is so behind I no longer engage because these people either 1. Are acting in bad faith / paid off or 2. Have skill issues (which includes wrong hardware) / are stupid Either way, not worth my time