Gerald Morrison

@0xSero The book cover on Amazon’s website is different so it’s a bit confusing. Maybe a newer edition? If it’s the same book, I will buy a copy. Thanks for highlighting.

3 books already gifted Why Machines Learn This will take you from noon to understanding the stack in a month. If you missed my giveaway one of the winners decided to pay it forward <3 If you love what I do you’ll love this book

@0xSero My job took me away from time to time. Coming home to the family created a sense of joy that can be had no other way except that it be preceded by absence. Keep up the good work “dad”! 🙂

I love traveling I love going new places. Since my son hit 2 years old, I have hated traveling alone. I miss him so much. I feel ridiculous getting so sad over a week away from my kid );

@0xSero First! That counts for something right? 🙂

Giving away 5 Opencode Go subs Winners selected randomly from comments in 24 hours.

@JuliaLaRoche @TruthGundlach @grok what are the top 10 key points and/or warnings in this interview?

The feedback on the @TruthGundlach episode has been incredible, so I'm posting the pod here on X too. 🙏 In his debut on The Julia La Roche Show, Jeffrey Gundlach, founder and CEO of DoubleLine Capital, breaks down why private credit is an unmitigated disaster, why the next recession will send rates up and the dollar down, and why most American investors are completely unprepared for what's coming. Timestamps: 0:00 Introduction & welcome Jeffrey Gundlach 1:33 Big picture macro: secular shift from falling to rising interest rates 16:00 The case for 100% non-US stocks 17:30 Gundlach's current asset allocation 22:00 Private credit and why it's a “total unmitigated disaster" 38:00 The Fed follows the 2-year Treasury - next move a rate hike? 42:30 Recession odds 47:00 Capital preservation mode: lowest risk positioning in DoubleLine's 17-year history 50:00 The gold call 53:00 The most dangerous force in investing 56:00 California headed for bankruptcy? 1:01:00 Non-consensus prediction: three parties on the ballot in the next presidential election 1:02:00 The Fourth Turning

@0xSero Got it. Thanks for setting up the Discord BTW. I’m looking forward to what you produce. 👍

@GeraldMorrison Different underlying engines and target audience. Lmstudio is llama.cpp and mlx only from my understanding. I am connecting to vLLM, sglang, exllamav3 I also want to have it be an agent launcher, I built this for myself and didn’t want to open source it but it saved me time

I am going all in on vllm-studio, in the past my take was that if Claude can do it out then people should figure it out. I've also just been doing whatever comes to mind, but I am going to trim out most of the code and focus on a desktop electron app. Good UX coming soon

We are quickly moving to the era of agentic AI. This will change the targeting protocol as the AI (one agent) makes the decision based on input from sub agents that each have a different validation task. This will make targeting decisions of higher quality without too much additional latency. That’s an improvement but it takes humans even further out decision making loop. With the required targeting speed, no human will ever be in the loop and it will be left to the quality of the agents analysis. Which means the quality of the input data. It also means that no one is to blame for mistakes because everything has been handed off to a complex system. And this can’t be stopped - it is an inevitable progression of technology used for military purposes. The school bombing is an unforgivable tragedy because higher quality data was available- it wasn’t used. In the future, a sub agent tasked with the responsibility to validate the correctness with the most recent information could have prevented it. I am not defending the use of AI for targeting decisions. That’s going to happen. But I am blaming those in the data extraction pipeline and those responsible for the guardrails in the AI decision-making loop. More powerful AI is coming quickly. The tools are there. We need to do better.

There is something deeply unhuman and evil about this system being built this way on purpose. Where compliance, not competency, is the end goal. Where by design no one is responsible and everyone has plausible deniability. Felt sick to my stomach reading that article. "when the US military and American manufacturers automated their factory floors, they consistently chose systems that were slower and more expensive but which moved decision-making away from workers and into management. The point was not efficiency – it was frequently extremely wasteful – but control. A worker who understands what they are doing can exercise judgment the institution cannot govern. Move that understanding into the system, and the worker has nothing left to do but follow instructions."

“Someone decided that deliberation was latency. Someone decided to build a system that produces 1,000 targeting decisions an hour and call them high-quality. Someone decided to start this war. Several hundred people are sitting on Capitol Hill, refusing to stop it. Calling it an “AI problem” gives those decisions, and those people, a place to hide.”

@profitsplusid @abcampbell @johnarnold Didn’t see your previous post. X Algo not the greatest for sure.

@abcampbell @johnarnold x.com/profitsplusid/…

I think I finally solved the stock market.

Awesome read - as usual. One thing I am thinking about with AI is that the top 10% of high income earners are white collar. They pay 50% of the tax. So when AI replaces them, the government gets slammed on tax receipts. Can the Fed cut when they need to entice the purchasing more and more T-bills?

I’ve been following Radigan for years. *** Just read this *** That’s the tweet.

The cost of memory alone is brutally expensive, as are the gpu cards. It’s hard to see a $1K machine running anything meaningful. What am I missing? I have 36 GB unified on a Mac and I am limited in the models I can use. I am always looking for ways to do more so I am very interested in this statement.

We will make frontier intelligence run on 1000$ of hardware by the end of the year.

@MagicCatCrypto @cgtwts 😂

@GeraldMorrison @cgtwts Yeah I buy Mac minis and put them in my closet and don't use them because I'm already paying 300 per month in ai subs....

RIP to everyone who just bought a $600 Mac Mini Base44 just dropped Superagent. > persistent memory > scheduled jobs > event triggers > browser sessions It’s basically an agent that runs your workflows on autopilot and is very simple to use.

@TheAhmadOsman Running a M4 Max, 36GB 512TB. Qwen 30B Instruct. LM Studio. Consistently getting 98-100 token/s when asking it to write a 200-300 word essay.

this is why i tweet about this another victim for the Mac Studio hype setting people’s expectations while not actually knowing what you’re talking about and convincing them to spend their hard-earned money because of it is straight up EVIL

@perplexity_ai @grok can this be run on my personal Mac Studio or do I need a separate Mac Mini?

Starting today, users can join the initial waitlist for the Personal Computer program. We will provide support and resources for the initial cohort. perplexity.ai/personal-compu…

Announcing Personal Computer. Personal Computer is an always on, local merge with Perplexity Computer that works for you 24/7. It's personal, secure, and works across your files, apps, and sessions through a continuously running Mac mini.

@itsolelehmann COWORK

i made a 3-day Claude Cowork for Beginners course, and it's yours for free by the end, you'll have a personalized AI teammate on your computer that: • knows your style • connects to your tools • and produces finished work you can send immediately here's what you get: day 1: install cowork, set global instructions, and run your first real task (15 min) day 2: workflows that replaced hours of my week, including building landing pages from a description and running full competitive analyses in one prompt day 3: skills, plugins, and connectors so cowork actually knows how you work and can access your tools + copy-paste prompts so you can follow along as you read like + comment "COWORK" and i'll DM it to you

@Scobleizer @OpenAI These reports are excellent - Thank you for posting!

Just added on an afternoon update to this @OpenAI GPT-5.4 report (new stuff at bottom). If you like these reports (all done by reading the entire AI community here on X) please let me know. If you hate them, let me know that too. :-)

I was holding off buying a Mac Mini but I pulled the trigger on a 24GB RAM, M4 Model to experiment. These models will be interesting. If I like where this is going, I'll sell the Mini and upgrade to a lower-end Studio (maybe). Thanks for highlighting these new light-weight models!

Do you understand what this means? Are you aware how much the world just changed? You can now run frontier intelligence on a potato Your $600 Mac Mini can now run unlimited super intelligence for free. No authoritarian AI companies can cut you off Do this immediately, no matter what device you’re on: 1. Download LMstudio 2. Find these models in the search 3. Look for the MLX ones if you’re on Mac 4. Download and load them 5. Ask your OpenClaw to use them for most tasks I thought the future was a year away. Nah. It’s today

@radigancarter Stay safe out there there!

Guys, don't worry, the war's on but I still got the letter out! radigancarter.com/p/the-hearts-o…

@Scobleizer This was a very helpful curation. A great use of AI. But will this tool be priced for the common man?

Anthropic vs Pentagon. The report of what X is saying about it. New kind of analysis and curation thanks to AI. docs.google.com/document/d/1EB…