Building chips | Founder @ Tendrils Compute
Joined April 2015
- Tweets153
- Following172
- Followers2K
- Likes460
Pinned Tweet
CPUs suck. We're building a new general-purpose chip that scales to thousands of cores while being more energy-efficient.
We're hiring hardware design engineers, consider joining us tendrils.co/jobs
What we do differently ...
We're hosting one of the most talent dense hardware dinners in SF
congrats to @nilscmr and the Tendrils team!!
we're hosting a dinner tomorrow w/ co-founders Nils & Hugo, as well as our great friends at @fractile_ai - DM if are interested in chip design and want to discuss the crazy future of computing.
3
47
2,258
Nils Cremer retweeted
congrats Nils and team Tendrils! enormously ambitious and technically special team — rarely do you see folks attach something as core as cpu architecture. thrilled to have you in the @conviction Embed family
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
We’re investing in British startup @TendrilsCompute through @UKSovereignAI, backing their ambition to reinvent the computer processor as part of a $30m funding round.
Turning British ingenuity into new industries, good jobs and growth across the country.
Our history tells us why it matters!
Nils Cremer retweeted
Tendrils is out of stealth
We have hosted many incredible thinkers at @theembassysf
@nilscmr is one of the early minds that stood out
It is rare to see someone that presents his vision to a Bay Area audience and be told that it might just be *too* ambitious, nonetheless agreeing that he was brilliant.
His response was to carry on in the same steadfast manner he always had.
Fast forward a few months after his stay with us, and the Valley had understood that this is a generational founder: His calm, intellectual clarity captivated every guest that met him at one of our dinners, or thereafter.
Very rapidly, it became clear to everyone that he is going to make the impossible possible.
Nils sees the future not as something aspirational, but something inevitable. As far as he's concerned, he is just doing the hard work to get there.
We're exceptionally excited to see Tendrils come to life, and grateful to be considered as part of the family.
(Also cannot wait for the next time we get to see him perform a magic show dressed as a wizard)
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
CPUs have been bottlenecked by technical decisions made in the 1970s for too long; modern workloads require modern architectures, and @tendrils is completely revolutionizing the CPU from the ground up. @nilscmr is incredible and I'm proud to be supporting him as an angel in this round with Benchmark. Let's go Tendrils!
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
Thrilled to see @TendrilsCompute raise a super strong seed round to build the future of CPUs for high intensity datacenter workloads. Congratulations on this milestone, @nilscmr and the whole team. An exciting future lies ahead for us all!
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
@nilscmr and team are absolute legends! Super excited to see lots of new chips and compute coming from Cambridge!
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
So thrilled to back @nilscmr and the team at Tendrils, who are taking advantage of advances in AI to rebuild the CPU.
Absolutely brilliant and ambitious founders who I feel lucky to be around.
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
The world needs so much more general-purpose compute. We raised some money to do that
Tendrils has raised $30M led by @benchmark to rebuild general-purpose computing. @jaltma is joining our board.
At Tendrils, we believe that the future of general-purpose computing lies beyond legacy architectures for legacy software.
Nils Cremer retweeted
It’s a good thing accredited investor rules exist to protect the everyday man from investing $5,000 in his friend’s grocery store so he can deploy that capital posting collateral to sell naked 0DTE gut strangles on AMD with margin
Nils Cremer retweeted
@nilscmr and @achterbrain and @DanAkarca is like a cartoon crossover episode. Very excited for what you all are building together.
There is an unsaturable demand for compute. Happy to be providing the significant increase we'll need
With our $100M seed round, we announce our first wave of global compute partnerships, flagship among them @CerebrasSystems.
Partnerships spanning silicon, infrastructure and geography. The path to intelligence that is orders-of-magnitude better, cheaper, and more sovereign.
CPUs treat memory as shared mutable state which means you always have to assume the worst case. You always need to be paranoid about someone else reading or writing your data while that's not the case most of the time
Nils Cremer retweeted
Today is Drop #2: Research.
Robotics foundation models need a lot of data, and a lot of engineering even for tiny adjustments.
This feels wrong.
Today, we share with the world a sneak peek to our research - training more efficient robotic foundation models.
1/3
Nils Cremer retweeted
100% of AI chip startups have slides/“simulated performance data” showing that their chip is way better, but 99% of custom ASICs fail. Why? The MATH isn’t MATH until you realize that AI chips are about Software. It is relatively easy to build a chip and put numbers onto slides; it is hard to build great software. That is why 99% of AI chip startups fail.