Are there any case studies of companies/cloud vendors using non-Nvidia GPUs for inference at scale?

Assuming startups like Etched (with its recent massive funding) could shrink CapEx quite a bit (and make it not such a large revenue shortfall)

Analysis of the "tidal waves" sweeping over the VC landscape. The post examines how some of the macro trends + rapid advancements in AI are reshaping the venture & startup ecosystem, which include:

- The impact of increasing fund sizes on returns (and why a $500M fund might need to generate $17.5B in exit value to 3X). - Novel strategies in which data science and AI can be applied in the VC investment process. - How VC, traditionally a “cottage industry”, is becoming more high-frequency

As well as some predictions on where the industry might be headed:

- How Solo GPs and smaller/nimbler firms could harness AI to rival much larger investment platforms. - The transformation of VC into a more traditional asset class (but with a twist!) - The potential re-emergence of ‘calm funds’ in a world of capital-efficient, AI-native startups - The changing role of CVCs and cloud hyperscalers in startup investing, and why massive funding rounds in foundation model startups probably won’t continue

Would love thoughts & feedback from the community!

Excited to share my thoughts on "model collapse". As the internet becomes increasingly filled with machine-generated data, I explore whether we are at “Peak Data”, how that might affect the efficacy of large language models & applications like ChatGTP going forward, as well as possible solutions that players at the foundation model & data layer might use to adapt to this new world.

What restrictions/things to know or consider are there for Canadian founders who are on a mix of TN/H1-B visas?

Those already in H-1B and TN status with their company or looking to obtain H-1B or TN status with their company?

Apologies -- this would be already in H1B OR TN status (at their existing company) and looking to start a company

I recently analyzed some fascinating financing trends as the US and China took very different paths in advancing their AI ecosystems (Part 1 of my Substack post [here](https://eastwind.substack.com/p/mad-china-and-the-semiconduc...))

In Part 2 of my Substack mini-series, I highlight why access to semiconductors is crucial to maintaining the pace of innovation in AI, and why up until the past several years, access to leading-edge onshore fabrication capabilities has been less of a strategic focus.

I contend that geopolitical tensions in Asia and our reliance on the supply of AI accelerators manufactured in Taiwan creates a dangerous reflexive loop, and how that affects the ability for companies in the space (like OpenAI) to continually innovate. I also explore some interesting areas of opportunity that entrepreneurs may look to exploit.

Would love thoughts & comments!

Since AI training is highly parallelizable one can look at it as a question of cost for training, that is given enough electric power, any nation state can train the model they want. And it would be nice to know how much cost reduction is due to fab process advance and how much is due to architecture and software advances. My suspicion is that the fab node is not strategic. It is a commercial issue to integrate more cores as fab advances, but training can be done over more chips and boards if integration isn't there, maybe for higher cost.

Hi all! I'm an early stage investor at FirstMark Capital.

In analyzing over 3,000 financings during my time working on the 2023 MAD Landscape (https://mad.firstmark.com/), I noticed some fascinating trends in how the US and China have taken very different approaches to advancing their respective ML/AI startup ecosystems.

In this two-part series, I explore how China has taken a much more “concentrated” approach to building its ML/AI ecosystem, and how this approach is designed to help China reach technology parity with the US (part 1).

I also highlight the importance (and the urgency) with which the US needs to achieve full semiconductor independence, so that the progress that we’ve enjoyed so far in the fields of ML/AI remain unimpeded (upcoming in part 2).

Feedback & comments appreciated!

Not sure about the "invisible" hand part. Note the difference could be explained by the structure of each respective economy. Chinese economy is not nearly as service centric as US (Chinese marginal cost of service labor still relatively low), so chatbot type of applications are not nearly as interesting to China, though that attitude could change with time.

Yup, very good point. With the publicity generated by OpenAI there's a heavy focus now to fund domestic competitors (see article published on the Information today, pay-walled -- https://www.theinformation.com/articles/sequoia-and-other-u-...).

The downstream effects on the application layer is an open question, however.

Hi all! First post & nearly decade long lurker here!

I recently built a data science system to help VCs (like me!) identify interesting open-source startups using the GitHub API. I wrote about my experiences & learnings here & would love for folks to check it out!