🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

2010s: data warehouse is the new backend 2020s: object storage is the new backend

The bear case in the GS report underestimates how much of day-to-day programming falls into the bucket of simple but tedious. Freeing developers from that work is precisely the point.

Okay, now we really have gpt4 at home

In-app search is a guessing game without a way to improve search results. Objective comes with query evals so you don't have to guess how well search works for your users.

Llama's permissive license that allows for distillation is the most consequential shift in the AI arms race. This is a neat example of lightweight prompt-distillation

One side effect of on-device inference: lose the body, lose the mind.

deploy AI search to prod in less than 5 mins with Objective, Inc.. as simple as upload JSON, create index, get results

Zero-cost inference is now even easier!

Two interesting bits from the GameNGen paper (diffusion models acting as game engines): 1) Data was generated with an RL agent — a clever way to get synthetic data. 2) Researchers used StableDiffusion 1.4 from *2021*. You don't need the next big model to do cool work.

Meet Anton. Latest release from Objective, Inc.. Improve search results with LLM as a judge. It works on any search you are already are using. Watch Lance Hasson demo it: youtube.com/watch?v=geX8mw…

Petition to return to company names like Software Arts Inc.

(WIP post) Sell aspiration, not efficiency

Current reflections on why I struggle to find some AI products compelling: sell aspiration, not efficiency

LM Studio 0.3.4 ships with Apple MLX 🚢🍎 Run on-device LLMs super fast, 100% locally and offline on your Apple Silicon Mac! Includes: > run Llama 3.2 1B at ~250 tok/sec (!) on M3 > enforce structured JSON responses > use via chat UI, or from your own code > run multiple