Life, Society, Technology: 🧵 1. On Things that really matter in Life !

Twitter is on a different timeline of its own where delusional larpers here burning 1.5B tokens with nothing to show somehow start thinking that these LLMs which have stalled in capability more than a year ago, are going to change the entire universe. Jokers all around

There is no next Amazon. It’s Amazon only. Ppl who think someone can build an Amazon competitor have really no clue about consumer nor what Amazon is.

A woman told this Asian boy on the bus not to stand next to her so he sat down in the only open seat & she freaked out saying she has thousands of followers of Instagram & “how dare you suspect me to this rage”. A sweet girl switched seats with him & was laughing at her. Racist?

I will take the other side of the bet. Agentic commerce is a super tiny TAM. And most ppl don't want agents to buy stuff for them. Give it a year's time

We turned Claude Code, Codex, Hermes, Pi, @opencode and other coding harnesses into RL environments. No changes to the harnesses, no changes to the training code. Any open model, any task set, fully open source my friends! Same model, same weights: 62% under Mini-SWE-Agent, 33% under Claude Code. But training inside a real harness normally means reimplementing it as an environment, so most models get trained in a scaffold nobody actually ships. The fix is a proxy, not a rewrite. The harness thinks it's talking to a model API. It's actually talking to a capture proxy that speaks the 4 formats coding agents use (OpenAI Chat Completions, OpenAI Responses, Anthropic Messages, Gemini), forwards to @vllm_project, records the exact token IDs and logprobs vLLM sampled, and hands TRL sequences it can train on. The harness becomes the environment. 10 harnesses run through it today, none modified. And because you control the reward, you can shape behavior the harness never asked for. We added a small bonus for solving a task in fewer tool calls: on tasks it already solved, the model now uses 31% fewer calls, in every harness, and about half under Codex. Tested on LFM2.5-2.6B from @liquidai: → Train in one harness: better mostly in that harness (OpenCode 34% → 58%). → Train in 4 at once: better in all 4 (42% → 54%). → SFT on 3,189 rollouts from Qwen3.8-27B instead: plateaus at 47.5%, below both RL runs. Everything is open and reproducible: the capture proxy in OpenEnv, the trainer in TRL, the tasks, the SFT data, the training code and all 7 trained models. Bigger models and bigger runs next. Full guide: huggingface.co/spaces/FineEn…

There is no bigger pump and dump artist that Brett Adcock

What happened to the marketing agents product that she was shilling last year but never launched

they are inventing indoors

RT @1kpier: next time they tell you AI art isn't art, show them this

Only in Silicon Valley

You are Welcome!

Confusing for crypto ppl because their IQ is anyway under 50. Most ppl don’t seek automation in life. They are not looking to save time. They are looking to waste time.

Wow After writing earlier that iMessage is the interface which can definitely be not used to spend time, now they claim Most ppl aren't looking to Save time, they are looking to spend time And Wabi thesis was opposite of this all this while Lol. Where did they learn from?

This gets more impressive when you realize he’s not standing

I haven’t tried Claude Code Desktop in a few days. 2 things Codex does better based on my experience: > better computer use. I use it to directly operate my computer for things like reimbursements. > easier to work with oss models, and generally less verbose.

Another insane Jev use case! Traditional database filters need precise, predefined conditions. But many questions are semantic: - Is this article mainly about software engineering? - Which topic best describes it? - How technically deep does it appear? These usually require moving rows into application code, invoking a model, parsing its output, and writing the result back. pg-jev is an open-source Postgres extension that exposes Jev through SQL functions. - jev() works inside WHERE - jev_prob() returns a probability - jev_choice() selects a label - jev_score() ranks rows across ordered levels. It needs no vector column or embedding index. Under the hood, pg-jev batches rows, sends them to Jev, and returns typed answers that SQL can filter, sort, group, and combine with exact predicates. The recording below runs it against real Hacker News stories stored in Postgres. It first shows the rows, then asks Jev to find software-engineering stories, classify them by topic, and rank them by technical depth. The final query reports requests, tokens, cost, and cache hits from the database session. GitHub repo: github.com/realZachi/pg-jev (don't forget to star it ⭐) Postgres keeps the data and controls the query, while Jev handles the part SQL cannot express as a deterministic condition. If you want to understand what Jev is doing under the hood, I also wrote a hands-on guide to building a Jev-style model with open models, entirely locally. Read it below.

😂 Why nobody leaves Twitter.

For someone who claims to write “thought pieces” you would expect to have a better thought response to this article than just blindly cheer lead it. Herd style mentality all over in valley TL:DR: Agentic Commerce is not an agent oppty. And won’t be

Flydubai captain Smit Machchhar: “I knew these might be my final moments, I was ready to die. But I told myself: just open the door, just do it.” One of the greatest heroes in history 🙏🇮🇳🇮🇱