swyx
swyx (@swyx) posts concise technical commentary on large language model behavior, with occasional references to OpenAI (OPENAI). Content is social-media–first and typically informal rather than detailed investment analysis.
Past bets that played out
No standout investment calls identified. Content consists of brief technical observations and links without actionable market or trade recommendations.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
What this channel is watching now
Mentions: OPENAI (1). Average conviction low (0.1). Current commentary centers on model behavior and the role of post‑training in instruction adherence rather than company- or market-specific claims.
Latest videos and market context
No recent video content on record. Recent posts are short text comments and links.
swyx @swyx 1m ok fine 4 new features - SmolForge is now getting customizable skins and spritesheet animations swyx @s...
A developer (swyx) mentions "SmolForge" adding customizable skins/spritesheet animations and dogfooding an "agentic GitHub clone" with built-in CI/CD "thanks to Workers for Platforms" before launch. This is an anecdotal early signal of developer adoption/experimentation around edge/serverless app platforms (most directly Cloudflare Workers/Workers for Platforms), but lacks concrete metrics, dates, or company attribution beyond the implied platform.
swyx @swyx 57m the reason i am working on a new gsuite is because of extremely stupid defaults like this bullshit 8 1...
A social media post complains about “extremely stupid defaults” in Google’s GSuite/Workspace and says the author is working on a “new gsuite” alternative. This is weak, anecdotal signal; at most it reflects minor dissatisfaction/competitive intent in the productivity suite market.
swyx @swyx 3h btw ive been dogfooding an agentic github clone over the past month or so and its gotten quite quite en...
A developer reports dogfooding an “agentic GitHub clone” with built-in CI/CD enabled by “Workers for Platforms” (likely Cloudflare). This is anecdotal but directionally supports a thesis that edge/serverless platforms are becoming credible devtool primitives and could expand into adjacent workflows (code hosting, CI/CD, agentic automation).
swyx @swyx 11h one thing i think people dont appreciate enough about @poolsideai is their unusual degree of openness ...
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Proof-backed call history
Recent source posts are brief social comments about LLM instruction‑following and post‑training effects, plus casual replies that include links. Posts do not contain concrete market, product, or financial claims tied to public tickers.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
Social post highlights PoolsideAI’s rapid iteration (“10k–20k experiments/month”), focus on coding models as a path to AGI, and unusual openness/open-model posture. No public-company financials disclosed, but it reinforces the ongoing buildout/competition in code-gen LLMs and the need for large-scale AI training/inference infrastructure.
swyx @swyx Sep 1, 2024 I saw an AI engineer today No cursor. No claude 3.5 sonnet. No aider. He just sat there. Codin... swyx @swyx Sep 1, 2024 I saw an AI engineer today No cursor. No claude 3.5 sonnet. No aider. He just sat there. Coding without running an LLM every other minute. Like a psychopath. swyx @swyx Sep...
A humorous anecdote implying heavy reliance on LLM coding assistants is becoming normal, but it contains no concrete data, product update, adoption metric, or catalyst. Actionability is very low.
A humorous anecdote implying heavy reliance on LLM coding assistants is becoming normal, but it contains no concrete data, product update, adoption metric, or catalyst. Actionability is very low.
A humorous anecdote implying heavy reliance on LLM coding assistants is becoming normal, but it contains no concrete data, product update, adoption metric, or catalyst. Actionability is very low.
A humorous anecdote implying heavy reliance on LLM coding assistants is becoming normal, but it contains no concrete data, product update, adoption metric, or catalyst. Actionability is very low.
Non-financial motivational hierarchy (“Systems > Goals … People > Projects”). No market, company, or asset-specific information; not directly tradable.
...ttraining for it @TheAnirudh i saw it in oai when it came out but i think this is all about posttraining for it @TheAnirudh i saw it in oai when it came out but i think this is all about posttraining for it Brief comment referencing "oai" (likely OpenAI) and that something is "all about posttraining"; no concrete market, product, or company details provided.
About this channel
swyx (@swyx) shares short, technical observations on AI model behavior. Posts are conversational and focused on engineering and research implications; they are not structured investment research.
@swyx
Most recognized assets
Unlock the full track record
Follow @swyx for concise technical takes on AI model behavior. Content should not be construed as investment advice.