Jay

Jay

[Jay @jma7889]

@MX3redux @JackPosobiec @grok you are evil

[Jay @jma7889]

@firesidealpha the most evil man alive

[Jay @jma7889]

@dnapway Of course China wants them. But they want unrestricted import with normal prices.

[Jay @jma7889]

@TimWarrenTrades @unusual_whales Without voter id, China can setup a billion signatures that not to reinstate

[Jay @jma7889]

@edliu1105 @nvidia @JensenHuang Don’t talk about Jensen as if he’s dead. He’s still here, and this doesn’t have to be the end. If NVIDIA employees truly stand behind him, speak up and push the board to bring him back.

[Jay @jma7889]

@dotey 不太赞同所谓学习AI帮不上忙的东西。 它都可以帮上好吧！要学会的是与AI共存合作。 不要放弃你已经掌握的知识。 要建立起更广泛的知识体系，放弃扣细枝末节。

[Jay @jma7889]

@Polymarket An Indian CEO found a new way to promote an Indian employee. The project will fail but will be considered as a success.

[Haider. @haider1]

Creator of C++, Bjarne Stroustrup:

AI-generated code isn't ready — it generates more bugs, more bloat, more security holes, and is nearly impossible to validate

"senior developers are already retiring rather than deal with it"

The problem is that even a small prompt change can shift the entire codebase in unpredictable ways

[Jiayi Weng @Trinkle23897]

Codex grew programmatic policies with no neural nets: max score on Breakout, and SOTA-level scores on MuJoCo.

Maybe heuristics were not too weak. Maybe they were just too expensive to maintain. Maybe it's the next paradigm.

https://t.co/1ZaIneleuW

[Jay @jma7889]

completely disagree. mcp solves one core issue of integration: interface mismatch, by adding a smart bidirectional proxy to existing apis.

[Maxime Rivest 🧙‍♂️🦙🐧 @MaximeRivest]

MCP was, technically, entirely unnecessary.

Psychologically, however, it is a masterpiece.

If a company makes an MCP, you feel like you have permission to use AI with that API service. It stimulates you to think: "Is there an MCP for that app?" or to install an MCP for that file system, when in fact you were always allowed to make a server on your computer, expose it to the internet, and have LLMs call it. It is mostly—only?—for that last point where MCP is slightly convenient. It helps chatbot apps (clients) know what internet endpoints they should let their AI use (not applicable to Claude Code, which already uses your machine to make any requests). But for a while, it was hard to get https://t.co/BQm7cbg8ts to make API requests, especially to endpoints requiring a key or other such authentication, and cross-browser issues were also getting in the way. But that didn't "require" MCP—it just greased the wheels a little.

[ℏεsam @Hesamation]

McKinsey studied 50 agentic AI builds and where they fail the most, and boiled it down to 6 key factors—essential for AI engineers:

1. It’s not about the agent, it’s about the workflow. don't obsess over building "impressive" agents. think about the whole system, not fun toys.

2. Agents aren’t always the answer.
Not every workflow needs a multi-agent system. Low-variance, predictable tasks are best handled with rules or ML, LLMs add complexity . The big wins for agents come in high-variance, messy processes (e.g. extract complex financial information)

3. Avoid "AI Slop". (common)
Focus on long-term development of agents, as you would with the development of an employee. Forget impressive demos. Double down on benchmarks. Agents should be given clear job descriptions, onboarded, and feedback so they improve regularly.

4. Track every step, not just outcomes.
Scaling agents up without visibility is asking for silent failures. Think about monitoring every stage of the workflow. This way teams detect errors early, refine logic quickly, and avoid total breakdowns. When mistakes happen (and they will), you can track where things went wrong and why. Don't skip this.

5. Reuse agents when you can.
Many companies waste time building one-off agents for each task. The smarter play is creating modular agent components (ingest, extract, verify, analyze) that can be reused for other workflows. Centralizing validated tools and prompts cuts 30–50% of redundant work, this number is no joke.

6. Humans remain essential, but in new roles.
Agents can parse, automate, and scale. But humans provide judgment, edge-case handling, and creative problem-solving. The future isn’t agent vs. human, but agent + human.

These are the mistakes startups and established companies make at scale. They cause massive damage to reputation and resources. And now you know how to avoid this.

[Alex Tomala @a__tomala]

Cute stickers i got recently :>

[Jay @jma7889]

Where AI is failing design systems, and where we are failing AI - https://t.co/bO61KMvWpu https://t.co/Go637RQJxl

[Jay @jma7889]

When claude code cli turned a 1 hour job into 4 hours struggle, then it's chatgpt5's turn. But claude code can't understand chatgpt5. At the end, codex completed the work in 30 min.  😢

[Jay @jma7889]

“Overall, we're seeing a ~30% increase in speed from the feature proposal to when it hits prod. This is huge for us.” 🥵

[Yuchen Jin @Yuchenj_UW]

This is how they vibe code at a FAANG:

>"get enough stakeholders to agree"
>"design review"
>"first few weeks of documentation"
>"PMs & TPMs breaking down tasks"
>3 months passed, then finally *vibe coding*!

What kind of hell is it to work like this?

[Jay @jma7889]

https://t.co/HZvhCbFOkx and https://t.co/B6FJO1sYK3

[Ben Lang @benln]

I put together a list of 30 startups who've each raised $50-100M in the past few weeks and are hiring:

[Jay @jma7889]

while hubspot, used to be an investors’ favorite ai software company, stock price cut in half last 6 months. content is king, seems true for ai marketing

[Jay @jma7889]

“Just this week, Reddit said its revenue was booming thanks to AI-powered ads. Pinterest shares, meanwhile, sunk this morning on the exact opposite— AI’s influence underwhelmedinvestors.” https://t.co/WrEFAdtgXQ

[Jay @jma7889]

Claude code spent an hour and gave up, for a one line change bug fix that eventually was fixed by OpenAI Codex in a minute. Claude was configured with context7 but Code wasn't. Better model seems still the answer.