trying to manufacture inference hw + sw @fdotinc @scale_at_gmi
San Jose/Taipei
Joined December 2021
- Tweets1.9K
- Following7.1K
- Followers1.3K
- Likes36.3K
Pinned Tweet
I love inference! Tap in🧑🏻💻
Welcoming @solidwillity to SCALE! 🚀
His startup, Touchdown Labs, makes AI inference optimization simple and accessible for resource-constrained organizations.
Can't wait to see what you build!
#SCALEatGMI #GMICloud #Startup
William Chen retweeted
if you think this way u r not meant to be a founder
i'm sure 90% of founders who raised real money would be better off financially just getting a job
we choose to build a company because there’s a chance to make the world a better place in a way that others don’t see
money is the byproduct
if you had 2 cofounders, raised from top teir vc, exited for a few billion dollars, you still wouldnt make as much as employee #400 at Anthropic will from the IPO.
Just let that sink in for a moment
William Chen retweeted
There’s no moat in software anymore and do not let anyone tell u otherwise.
All software can be cloned now
Starting to see “AI will eat software” predictions slowly coming true.
My barber just moved his booking system off booksy after 5 years because a friend of his learned about Claude and vibe coded nearly the same thing for his website in a weakened. He told me it makes no sense to pay booksy anymore.
I think this trend might accelerate starting from low margin and mom-and-pop businesses then spreading.
William Chen retweeted
Never, ever let them tell you it can't be done.
Palmer Luckey says analysts claiming it's too expensive to manufacture in the US are lying to protect megacorps
"I think I have a more realistic view than almost anybody. One of the problems with all these analysts that are telling you that it's going to cost $37,000 to build an American iPhone or $100,000 to buy an American computer."
"They're trying to prop up a status quo that the megacorporations want to keep going. They want to outsource everything. They want to hire cheaper non-American labor."
"They are willing to say and do anything to convince you that it can't be fixed, that it can't come back, that America is this helpless, neutered, post-relevant empire, and it is just not the case."
"How do I know that? Because I have to make stuff in America every day because I work in the defense industry. The stuff that Anduril makes every day is not actually that different from a computer. My drones are basically computers. They have memory, a GPU, a CPU, sensors, batteries, a rugged integration."
"The only difference between an Anduril attack drone and a laptop is a keyboard and a screen. That is not that hard to figure out, especially for a guy like me who started Oculus and comes from the consumer electronics space."
"We're not going to let people lie about what America can and can't do anymore. We're just going to do it."
William Chen retweeted
>be benjamine liu
>immigrate from taiwan to california at age 2
>grow up in thousand oaks
>go to yale
>study computational biology at cambridge
>become a rhodes scholar at oxford
>use AI to research alzheimer’s and parkinson’s
>try pitching new drugs to pharma companies
>get rejected
>“we have more drugs than we can afford to develop”
>realize drug discovery isn’t the bottleneck clinical trials are
>everyone is using AI to discover new drugs
>bet on developing them faster instead
>cold email legendary investor michael moritz
>get a reply in 28 minutes
>get a $2.25M check
>cofound trialspark in 2016
>rebrand to formation bio
>partner with openai and sanofi
>raise $372M series D led by a16z
>build a $1.8B AI pharma company
>raise over $600M to make medicine move faster
Today we’re announcing the $600M Formation Bio Drug Development Accelerator Fund.
@lintaho and I started @FormationBio because we believed the biggest bottleneck in bringing new medicines to patients wasn’t drug discovery, but drug development.
As AI dramatically accelerates discovery, that bottleneck is only becoming more acute. Without the capital and capacity to advance promising drugs through clinical trials, much of that progress won’t translate into medicines for patients.
Our new Fund brings together dedicated capital, drug development expertise, and AI-powered development capabilities to acquire promising drugs and advance them through the clinic.
William Chen retweeted
Palmer Luckey says analysts claiming it's too expensive to manufacture in the US are lying to protect megacorps
"I think I have a more realistic view than almost anybody. One of the problems with all these analysts that are telling you that it's going to cost $37,000 to build an American iPhone or $100,000 to buy an American computer."
"They're trying to prop up a status quo that the megacorporations want to keep going. They want to outsource everything. They want to hire cheaper non-American labor."
"They are willing to say and do anything to convince you that it can't be fixed, that it can't come back, that America is this helpless, neutered, post-relevant empire, and it is just not the case."
"How do I know that? Because I have to make stuff in America every day because I work in the defense industry. The stuff that Anduril makes every day is not actually that different from a computer. My drones are basically computers. They have memory, a GPU, a CPU, sensors, batteries, a rugged integration."
"The only difference between an Anduril attack drone and a laptop is a keyboard and a screen. That is not that hard to figure out, especially for a guy like me who started Oculus and comes from the consumer electronics space."
"We're not going to let people lie about what America can and can't do anymore. We're just going to do it."
William Chen retweeted
lot of people making stuff up so here's some more context
you can imagine there's hidden enterprise demand for open source models that isn't visible
but if you look at the top inference providers that aggressively go after this business, they have shared some numbers and they are all roughly the same size as us. and at best there's 5 of them
if you think enterprises are self hosting inference on their own GPUs, there are so few people who can run reliable inference and they get paid a lot and work at ai labs. and if they really want sovereignty they're not getting that by leasing GPUs from a third party
the root point here is this is entirely the wrong way to look at things, you just have to look at products
how many products exist with a lot of traction that can use open models? this is what creates demand, it's not some magic thing that grows on its own
OpenCode is one of those and we do see a lot of enterprise inbound. but it's so early, all of these companies are just starting to look at mixing in some OSS inference
so yes demand will probably go up in the future but everyone in this industry is investing resources in preparing for that future. and so few people are building the products that will make it happen
William Chen retweeted
🚨🚨 There will be no memory oversupply, even beyond 2030
As you all know, HBM demand is enormous, and meeting that demand is constraining the production capacity available for conventional DRAM. HBM already requires far more wafers than conventional DDR5 to produce the same number of bits.
So what happens when new 3D stacked DRAM architectures, integrating DRAM directly with logic, begin to scale around 2029 or 2030?
Could that put even more pressure on conventional DRAM supply? As far as I know, the memory companies have not even finalized which DRAM process nodes they will use for these architectures. And if my understanding of the architecture is correct, 3D stacked DRAM could impose even greater constraints on bit output than the DRAM used in today’s HBM.
I think many analysts still aren’t even considering the supply constraints from 3D stacked DRAM. We’ll probably only start hearing about this next year.
Many companies in the industry are already collaborating with memory vendors on 3D stacked DRAM. These are things I already discussed in my articles two months ago, but I’m bringing them up again.
With SK hynix posting strong results, this seems like a good time to revisit the topic 😊
William Chen retweeted
🚨🇨🇳 A Chinese glass company just turned the RIOTS in France into FREE advertising!
Hengxinfeng filmed NOTHING, just posted a video of its unbreakable windows on social media, and went VIRAL.
EPIC TROLLING!
William Chen retweeted
important for all to note:
1 - venture returns are predicated on speed + scale
2 - reinventing a factory for speed + scale IS NOT "i need a factory to make x"
3 - venture equity is the only thing you can use to reinvent a factory
4 - reinvent it, then scale it with equity + debt
5 - scaling laws here far surpass most software companies especially on longer timelines, enormous cash flows and growth potential
For people who need pcb
made a tool for searching for PCB assy/fabs throughout the world: zackslab.com/atlas
if you're favorite pcb house isn't on here lmk :)
William Chen retweeted
the chinese are asking photonics PhDs to give training class to prisoners and make them do some of the assembling work, how are you going to move optics/photonics supply chain back to the US, calling a 4 pm meeting "in the evening"?
William Chen retweeted
American glass. American metal. American silicon. American batteries. American workers. American everything.
BlueChip is the first American computer in decades. Don't let anyone lie to you and say it can't be done.
Designed in California. Built in America.
Bc nobody uses AI except for startups in proportion to the 8 billion people in the world and that’s active usage not including ambient usage with agent swarms and sandboxes
Intelligence will and should become a commodity like water and for the most part electricity in the first world
Too cheap to meter: It is 0.025 cents (not dollars) per million tokens.
Look at the insane performance numbers for Kimi 3.1 on Vera Rubin
- 90$ worth (45M tokens * 2$ per million for input, assuming most tokens are input) of Kimi 3.1 tokens for 1$ of TCO at 100 tokens/sec interactivity. Insane margins. Lot of scope for players to cut prices.
It is true: The more you buy, the more you save 🙇♀️
- With theoretically same uplift for DS v4.1 Flash on Vera Rubin from Grace Blackwell (1600M per dollar TCO), there will be nearly 4000M tokens per dollar of TCO.
It is 0.025 cents (not dollars but cents) per million tokens.
(Grace Blackwell chart below for comparision. Then look at SemiAnalysis chart).
Explain me again how there is going to be 250B of inference GPU rental revenue next end year (25GW of inference ompute), and even higher GPU revenue year on year to justify 1T USD of per annum investment that will go on in perpetuity.
It is nearly 30$ per living person of GPU compute (8 billion population) and 120B tokens per living person (babies, kids, old people included) at 0.00025$ generation cost for a million tokens of DS v4.1 quality.
That is the level of consumption we need.
Either we have a glut or everyone has AI waifu, BF, dad, mom whatever they desire. I don't see everyone having a personal robot before 2027. I don't know about you.
Let me know if you disagree. These are relatively idealized scenarios. There would be lot of slack in the system and generation may be few tens of billions per living person, but even then it is a lot!!!
love this
PCBs made in America. Twice as fast as Shenzhen.
Today @CircuitHub is unveiling The Grid, our factory-scale robot in Massachusetts that builds circuit boards on demand.
No changeovers, no per-design setup. Prototypes and production batches, fully automated.
William Chen retweeted
Frontier labs making custom chips is not really as hard / as impressive as you might think.
1. They are primarily targeting inference not training. inference is easier
2. They are targeting transformers only. Think large MXU matmul units
3. They are writing all kernels in close to ISA with coding agents RLed for their ISA.
4. They aren’t making high level compilers for codegen
A frontier lab vending a true generic chip like a Nvidia GPU or Google TPU is MUCH harder than spinning up an in house inference program.
Still I think you will see many more inference chip companies as the cost of the software is just the cost of tokens.
William Chen retweeted
Western team have independently verified that DeepSeek's performance numbers of 38$ dollar revenue per GPU at DeepSeek's rocket bottom prices is true.
This is a momemnt when you update your priors. At this rate with just 1 GW, you can make a revenue of 150B per year.
Even with lot of slack it would be 100B per year. If you use half the compute to come online by 2027 end, you will have 25GW for inference with potentially 2.5T revenue at DeepSeek rates and 10T at USA lab rates (world GDP is 125 T, 10T is 8% of it)
You answer for your-self are we overbuilding!!!
vLLM matches DeepInfer, DeepSeek's own engine for DeepSeek-V4.1-Flash, on DeepSeek's decode configuration across batch sizes 8 to 256, and does it at DEP32 against DeepInfer’s EP64DP64.
DeepInfer was originally built on vLLM. By integrating DeepSeek's open-source kernels and extensively optimizing prefill and decode, vLLM now matches DeepInfer's performance curves.
William Chen retweeted
Today I had in-depth conversations with optical, packaging, and system design engineers at Intel, NVIDIA, Marvell, Micron, Broadcom, and Ayar Labs.
My takeaway is that optics is just getting started. Don’t put a ceiling on how big it can get.
Investing is still hard. Conversations like these strengthen my conviction to keep going.