Avui hi pensaré

l'infern està empedrat de bones intencions

[Avui hi pensaré @avuihipensare]

12/ No por esto, dejo de amar de donde soy, de mi país, de mi cultura. Sin embargo cada día más afirmó soy un amante de las culturas, de lo países y sus personas.
Espero nos volvamos a ver pronto Tarragona

Moltes gràcies @universitatURV @Red_Carolina y Tarragona

[Avui hi pensaré @avuihipensare]

1/
¿Qué cosas nos hacen humanos? el hecho de sentirnos que pertenecemos a algo, el amor, al final de vivir.

Soy y vengo de una tierra lejana, al llegar a este territorio desconocido, lo que menos me imaginé fue volver a nacer.

[Avui hi pensaré @avuihipensare]

11/
Sin importar, hoy solo me gustaría estar allá e ir a sentarme en un banco y ver como la gente pasa o quizás ir al comedor y hablar con las personas.

[Kin @KinGao476942]

卧槽！兄弟姐妹们，想用 Opus 5.5 做视频，一定要收藏这 8 个 GitHub 仓库。

1. awesome-opus5-5-videos（约 7100 star）
目前最大的爆款视频合集，每条都附上了背后的提示词。
https://t.co/VWhSZgNz6C

2. lemo-opuscar（约 4400 star）
一个 Claude Code skill，装上之后挑一种风格，就能做出纯代码短片，51 种风格随便选。
https://t.co/CSj7tyu8ie

3. pdoom-video（约 3000 star）
一支歌词卡拉 OK MV 的完整工程，本地就能渲出 4K60。
https://t.co/3WbzuFefAL

4. PDoomVideo（约 2300 star）
同一首歌的另一支 MV 源码，和第 3 个对照着拆，很有意思。
https://t.co/xnriTUB8nQ

5. mg-styles-15（约 1400 star）
15 种 MG 动效风格，成片、提示词、源码全都给。
https://t.co/LkRmgFXx7b

6. ClaudeAnimationBase（约 1300 star）
手绘卡通角色动画的起步包，自带 31 种表情。
https://t.co/wkXHp8xffq

7. awesome-claude-5-5-videos（约 560 star）
视频工作流合集，每个案例都标了出处，Opus 的就有 301 个。
https://t.co/uOw55tL0CQ

8. opus-video-prompts（约 510 star）
中文案例，配上能直接复制的提示词。
https://t.co/1WH40CnNub

[Avui hi pensaré @avuihipensare]

focus and discipline as a key for moving

[Puneet Patwari @system_monarch]

13 years in backend. Principal Engineer at Atlassian.

Concepts I wish I'd learned in year 2 instead of year 8:

→ p99 latency, and why averages lie
→ Connection pooling
→ Backpressure
→ Idempotency keys
→ Reading query plans (EXPLAIN ANALYZE)
→ Write amplification
→ Consistent hashing
→ Change data capture
→ The outbox pattern
→ Tail latency amplification

None of them are hard.

Every one of them would have saved me from tonnes of headaches.

[Alba @albacnt]

🚨 ADIÓS A ELEVENLABS

Un desarrollador ha creado una alternativa open source para clonar voces, doblar vídeos y generar audio desde tu propio ordenador.

Se llama VoiceStudio y es una locura.

Te permite:

→ Clonar voces desde un clip de referencia
→ Doblar vídeos a otros idiomas
→ Generar audiolibros y locuciones
→ Transcribir audio a texto
→ Elegir entre múltiples motores de texto a voz

Todo desde una misma herramienta.

Y lo mejor: puedes ejecutar los modelos compatibles localmente, sin depender de una API externa para cada generación.

El proyecto está en GitHub.

Guárdate esto, vas a querer probarlo.

Link abajo 👇

[Mansi 👩‍💻 @MansiCodez]

Go GOD LEVEL in DISTRIBUTED SYSTEMS by building these systems NOW: 🧵

[momoiika @moiiikaaa]

My biggest fear as a dev is AI disappearing and I have to sit to learn CSS again.

[Anurag @anurag_gharat]

Here are the resources that I am using

Videos:
1. https://t.co/5mV3fmHUkG
2. https://t.co/xkfUZIXXnw
3. https://t.co/RjMpmorzoO

Blogs and Articles:
1. https://t.co/tWMKocIuLr
2. https://t.co/0j3cGfYJlj

[Suraj Sharma @suraj_sharma14]

If you learn AI Infrastructure Engineering, you will never be unemployed again.

Stage 1: Systems & GPU Foundations
- Learn: C++, Rust, CUDA memory hierarchy, PCIe vs NVLink, thread blocks and warp scheduling.

- Practice: Write a custom CUDA kernel for matrix multiplication from scratch and profile it against cuBLAS.

- Why: You cannot optimize what you do not understand at the silicon level.

Stage 2: Transformer Inference Physics
- Learn: Prefill vs Decode phases, arithmetic intensity, memory bandwidth bottlenecks, KV-cache math.

- Practice: Profile a HuggingFace model with Nsight Systems to identify the exact memory bottleneck during token generation.

- Why: LLM inference is almost always memory-bound, not compute-bound.

Stage 3: Modern Serving Engines & Batching
- Learn: vLLM, SGLang, PagedAttention, continuous batching, chunked prefill.

- Practice: Deploy a 70B model and tune chunk sizes to maximize throughput without starving decode requests during long-context prefills.

- Why: Naive batching leaves 60% of GPU VRAM wasted. PagedAttention fixes this.

Stage 4: KV-Cache & Memory Optimization
- Learn: Prefix caching, KV quantization, CPU offloading, multi-turn cache reuse.

- Practice: Build a routing proxy that directs requests with identical system prompts to the same replica to share KV blocks.

- Why: Reused cache is free speed. It can slash Time-To-First-Token (TTFT) by 80%.

Stage 5: Quantization & Compression
- Learn: FP8, INT4, AWQ, GPTQ, Sparsity, TensorRT-LLM calibration.

- Practice: Serve a model in FP8 vs FP16 and benchmark the exact perplexity drop against the latency and VRAM gains.

- Why: Quantization is the only way to fit frontier models on edge GPUs and protect your margins.

Stage 6: Kernel-Level Engineering
- Learn: Triton, FlashAttention, CUDA graphs, operator fusion.

- Practice: Write a fused Triton kernel for RMSNorm or Softmax and benchmark it against the native PyTorch implementation.

- Why: Python overhead kills inference. Fused kernels save milliseconds that compound at scale.

Stage 7: Distributed Inference & Parallelism
- Learn: Tensor parallelism, Pipeline parallelism, Expert parallelism (for MoE models).

- Practice: Shard a 405B parameter model across 8 nodes and measure the communication overhead between GPUs.

- Why: Single-GPU inference is dead for frontier models. You must master the shard.

Stage 8: Speculative Decoding
- Learn: Draft-target models, Medusa heads, acceptance rates, n-gram drafting.

- Practice: Build a pipeline where a local 3B model drafts tokens for a cloud 70B model to verify in parallel.

- Why: 2x decode speed at zero quality cost is the closest thing to a free lunch in inference.

Stage 9: Multi-Node & Hardware Interconnects
- Learn: NCCL, RDMA, InfiniBand, NVLink, Disaggregated Prefill/Decode architectures.

- Practice: Set up a multi-node cluster and profile the network latency of tensor parallelism across nodes vs within a single node.

- Why: Network latency is the new GPU bottleneck. Disaggregating prefill and decode is the 2026 meta.

Stage 10: Cluster Orchestration & GPU Scheduling
- Learn: Kubernetes GPU operators, Ray, Slurm, MIG (Multi-Instance GPU) partitioning, KEDA.

- Practice: Build a queue-based autoscaler that spins up spot GPUs based on pending inference requests and drains them when empty.

- Why: Idle H100s burn $3+/hr. FinOps and scheduling are now core infra responsibilities.

Stage 11: AI Gateways, Routing & Observability
- Learn: TTFT/ITL SLOs, semantic routing, DCGM metrics, OpenTelemetry for LLMs.

- Practice: Build a gateway that routes simple queries to a quantized local model and complex reasoning to a frontier API based on prompt complexity.

- Why: Routing protects your margins and DCGM metrics tell you when your GPUs are silently throttling.

Stage 12: Public Benchmarks & Teardowns
- Learn: Reproducible methodology, latency/throughput Pareto curves, cost-per-token analysis.

- Practice: Publish a teardown comparing vLLM vs SGLang vs TensorRT-LLM on your specific hardware with full configs.

- Why: Public proof of hardware mastery gets you hired instantly by top AI labs.

Wrappers are a commodity. AI Infrastructure is the physics of scale.

The modern AI Infra Engineer builds the silicon nervous system for global intelligence.

Bookmark and Repost!