FOLLOWS YOU. Artificial Intelligence, Automation & Optimism. Everything I say is 100% serious...
San Diego, CA
Joined July 2022
- Tweets18.2K
- Following887
- Followers38.2K
- Likes24.8K
I've just started a swarm with Gemini 4 Carbon to build "Factorio AI Deathmatch"
basically PvP factorio... for large language models
my desktop now sounds like it's about to take off, there's several factorio processes running in the background and ambient temperature increased
7
13
1,216
Runway demonstrated an electronic flute project using Claude Fable 5.2, Runway and Blender.
I like this direction. Being able to develop a design across several tools could make experimenting much more practical. You could try different shapes and finishes before spending more time on the version you actually want to build.
6
8
1,323
The New York Times reports that Jensen Huang had an extramarital relationship with Nvidia healthcare vice president Kimberly Powell, who reported to him.
The allegation comes as Nvidia’s board investigates whether Huang was targeted by a foreign intelligence operation. The company has appointed outside counsel to conduct that inquiry.
3
6
1,371
It’s hard to picture Nvidia without Jensen Huang. WSJ now reports that the board is preparing to remove him.
If that happens, I’d be watching the succession plan closely. Replacing a founder so closely associated with the company’s direction would be an enormous test.
Nvidia sits at the center of the AI industry. How it handles a leadership transition could matter to customers and partners far beyond the company itself.
ByteDance co-founder Zhang Yiming reportedly told staff that matching OpenAI and Anthropic at the frontier would take until at least 2029 or 2030.
According to Bloomberg, he said Seed’s pretraining datasets are orders of magnitude smaller than those of frontier competitors and urged the team to avoid relying on distillation.
A Muse-like assistant is planned for the coming weeks.
Google DeepMind’s AlphaGenome 2 looks like a meaningful step for disease research.
In one example, it reproduced a known inflammatory bowel disease finding: a DNA variant increases activity of the ETS2 gene specifically in certain gut immune cells.
That level of detail matters. Knowing a genetic change is associated with disease leaves a lot unanswered. Pinpointing the gene it affects and the cells where it acts gives researchers something much more useful to investigate.
This is where I find AI in biology especially promising.
1
6
872
Grok Bot 1.0 is going after the work that falls between meetings.
Shared Bots can work with a whole team in Slack and Teams, remember decisions and follow up on unfinished tasks. They can also help produce documents and presentations.
I can see the appeal here. You agree on something in a call, and the assistant already has the context to help move it forward. Getting that continuity right could save people a lot of repetitive explaining.
7
16
1,191
Singulus will supply SpaceX with a solar-cell production line valued at more than €60 million, with delivery scheduled for the first half of 2027.
The equipment will support Starmind satellite manufacturing in Texas.
SpaceX has ordered 100 GW/year of HJT solar manufacturing equipment from Singulus in Germany.
This is for Starmind to create a sentient sun.
3
19
1,365
jobs give resources, meaning and status
those go away if AI replaces you
we need to
1) implement a plan to be able to have resources without a job (urgent)
2) let AI steam roll on through
3) figure out the meaning and status stuff later, it's not urgent
the resource plan MUST not be leave room for misunderstanding or corruption down the road
governments cannot take it away no matter how "bad" they deem a person to be
no one can loan against it, all such loans become uninforceble
basically everyone gets "f*** you money" that can't get taken away
let's say you need $50k to survive and you are guaranteed to get that for life
you are now your own boss, free from financial obligations
but we can't pay everyone that right?!
yes, but we are paying right now in US about $4k per capita in various welfare (higher for certain states)
a BIG chunk of ALL costs are human labor
if AI automation cuts costs over time, the minimum needed to survive might go from 50k to 25k to eventually being that 4k that we are already paying
we don't need more money, we need less cost
this is the future if we don't screw it up
say it with me:
let the robots do the work
Alibaba opened Qwen4-Max in preview, targeting coding, knowledge work, frontend development and complex agent tasks.
The model supports text, image and video inputs, a 1M-token context window, tool calling, web search and structured outputs.
API access is available through Token Plan, with early access also offered in QwenWork, Qoder, the Qwen app and QwenStudio. The API preview runs through November 25, giving developers a window to evaluate it inside their existing workflows.
Qwen4 is coming.
Deeper knowledge. Richer world understanding. Stronger coding and agentic capabilities.
Built for real-world tasks — from complex coding and knowledge work to front-end builds and professional software workflows.
Qwen4-Max is now in preview on Alibaba’s Token Plan. You can also get early access to the new model on QwenWork, Qoder, Qwen APP, and QwenStudio.
Put it to the test on a real problem. Show us what you build. 👀
Token Plan: qwencloud.com/pricing/token-…
interesting read, I think it misses a few points about things like curation and distribution, but very interesting nevertheless
overheard from friends talking about AI in China, and it’s very alarming
- AI short films went massive in China market because of how good the video models are now
- the first wave of people who mastered using AI to generate short films and comics made quite some money because they could produce a lot more content than traditional workflows
- but within just a few months, everybody and their mom could make equally good short films because all the expertise could be written in a markdown file
- when everyone’s more productive, no one actually has an advantage and it’s a race to the bottom. very quickly it became impossible to make money, and everyone’s busier than they were without AI
- the only one making money now is companies like bytedance who owns user demand and distribution. they sell ads to creators who pay to get traffic, and eventually ad spend became 80% of those creators’ operating expense, while AI tokens is less than 5%
so what are the problems here?
1. there’s no moat if you are creating anything with AI
if a human is really good at something, and it’s hard to find more humans who can do the same, they can leverage the scarcity of their skill to get paid very well
but if AI is good at something, it will be infinitely replicated, so there’s no scarcity hence you can’t really expect a sustainable business by creating anything with AI
2. more creators created more content people actually liked. that’s a good thing
but the economical benefit of that all got absorbed by the distribution platforms
there’s a famous chinese poem 遍身罗绮者，不是养蚕人 - it basically says people who dress well aren’t the people making dresses. that’s exactly what’s happening here
if our society doesn’t figure out a better way to distribute wealth, AI will help billionaires become trillionaires while average people become more exhausted than ever
2
13
2,613
imagine anyone can produce an amazing videogame and so 1,000 masterpieces are released daily
what does that mean in reality?
first, you rely on curators whose taste you trust to tell you about the things that are the best of the best and match your taste
(word of mouth can be part of this or a standalone category)
two, distribution, places like Steam manage your game collection and releasing on that eco system gets you more eyeballs than if you just put up a download link on the internet somewhere
when it took a studio to produce a talk show, a few people made all the $$$$
now anyone with a webcam can produce it. the people at the top still make the bulk of $$$$
(80/20 aka paretto principle stands no matter what)
there's more more of a long tail, more people trying but not really making anything
but the point is that the worst performing podcast on the planet has access to the same equipment as Joe Rogan, everyone can post on YouTube, Spotify, iTunes etc
so what?
the printing press made it so we all can get access to books, everyone learned to read/write, so authors can't make money after the printing press?
JKR pulled in billions with Harry Potter
did she have access to special letters or word combinations that the rest of us don't?
I bet you know how to write all the words in Harry Potter, I bet over your life you wrote the equivalent number of words as Harry Potter, but earned less than a billion from it
(true for *most* of the people reading this)
the idea that if everyone has access to some technology then everyones output and results are the same is complete BS
scribes should not be the only ones that can write
hollywood execs should not be the only ones to be able to produce money
software engineers should not be the only ones to produce software
doctors should not be the only ones able to work on curing cancer
mathematicians should not be the only ones to be able to work on math
every time any technology breaks one of those barriers, the world becomes a better place
the ONLY problem we have right now is that AI is about to break all those barriers at once and we do need some transition plan for those displaced
(again, most of us)
if we don't people get scared and when people are scared they do and say stupid things
things like:
"AI should not be allowed to do math, cure cancer, write code, make pictures..."
and
"let's ban AI from doing those particular things"
jobs give resources, meaning and status
those go away if AI replaces you
we need to
1) implement a plan to be able to have resources without a job (urgent)
2) let AI steam roll on through
3) figure out the meaning and status stuff later, it's not urgent
1
3
398
Google released Gemini Decision 1 and the Decisions API for tasks such as selecting tools, routing requests and checking rules.
It can evaluate up to 1,000 independent questions against shared text, images, audio, video or PDFs within a 256K-token context window. Results include probabilities and confidence scores for yes/no, multiple-choice and rubric-based questions.
Pricing is $0.10 per million input tokens, with no output-token charges. A free tier is available for prototyping, giving developers a way to test these decisions inside their existing workflows.
Google Labs introduced Tabletop, an experiment that turns your group’s links, files and ideas into shared interactive tools.
Ask @tabletop to build a road-trip map, tournament bracket, voting board or seating chart, then invite others to use the space together in real time. Starter templates are also available.
The useful part is combining the planning material with tools the group can actually interact with. Access is rolling out through a waitlist for US users aged 18 and over.
5
872
where has "throwing Compute at it" NOT worked?
specifically do we have ONE example where the goal is clear and throwing more Compute at it gives ZERO results?
(also why did my phone start capitalizing Compute all of a sudden...)
Really terrible idea to make predictions like this if you want people to take AI, including AI risk, seriously. A lot of medical problems are fiendishly complex and not just a matter of throwing enough compute at them. In some cases what we view as one disease is a bunch of 'em!
2
20
1,393
MiniMax released H3-Image, an open-weight model for image generation and editing. The company says it can generate native 4K images in around 10 seconds.
It supports up to 16 reference images and is available through the MiniMax API, MiniMax Code and Hailuo AI, with weights on Hugging Face.
High-resolution output and multiple references could be useful for building visuals around existing assets. One practical limitation: the API currently doesn’t support masked edits or transparent backgrounds.
6
15
1,604
Stripe has completed its acquisition of OpenRouter, following the agreement announced on August 19.
At the time of that announcement, OpenRouter reported processing more than 10 trillion tokens daily across 400+ models, serving over 10 million developers and companies.
OpenRouter says its product and existing integrations will continue, with model selection remaining neutral.
6
19
1,534