Investing @ Lionschain
Brooklyn, NY
Joined March 2015
- Tweets159
- Following3K
- Followers429
- Likes2.6K
Niftical retweeted
BREAKING: Revolut moved ~2,000 engineers onto self-hosted GLM for coding and pushed Claude Code and Codex to the exception list.
Per David Abitbol, who leads AI ops there: 56 Nvidia B200s, a blended 30 cents per million tokens, roughly a tenth of what they paid Anthropic and OpenAI. Engineering lead time improved 70%.
Here's the part that matters: the gains showed up when tokens stopped being rationed. Engineers quit saving AI for the hard tickets and pointed it at everything.
Frontier models become the exception, open weights become the default. That's a pricing problem for every lab heading to IPO.
h/t @OliverMolander
Would you move your team off frontier APIs if the bill dropped 90%?
Niftical retweeted
What if I want to use the RAM to do things that are prohibited by the terms of service of whoever is running the datacenter?
Niftical retweeted
Heard of a top law firm completely cutting off Harvey and all API calls to Claude/OpenAI after they saw what openAI did with navier strokes
Now hosting everything locally with GPUs they own, running Kimi and deepseek
Think you’ll see a lot more of this going forward
Niftical retweeted
The team at @actualinc has had early access to Nemotron 3.5 Super VL and used it to show phenomenal performance of the model on DGX Station using actual. It’s been a privilege to work with them these past few months. Their blog is cross linked in the official launch blog 🚀💚
Actual Computer turns your NVIDIA DGX Station into a department-scale AI factory.
Read the full case study:
actual.inc/company/blog/case…
yes - this is always true on some level but also is the antithesis of the goal of most VC backed startups and what Buffet teaches and what PT describes in 0-1 etc... the goal that all investors and founders strive towards is for the business to have a moat so defensive that a monkey could run it, and as we've seen with facebook, uber, ebay, etc. such a moat is possible, because they have all at some point suffered egregious failures of governance and yet all are enormously valuable and durable businesses. this is the dream and also the basis for tech companies to be worth trillions.
none of this is to say that you can't build a great business on the back of continuous innovation. it has been done many times before and will undoubtedly be done many times again. but it is much harder to do well and harder still to sustain over long arcs of time. there are exceptions of course but they are few and far between.
as @nikitabier stated (much more concisely) - in today's world AI has removed much of the barrier to developing new products and thus the avenues by which moats can be established are arguably fewer than they have been in the past. going forward, many of the businesses who drive unique and differentiated outcomes will do so by generating and leveraging proprietary data and imo their ability to maintain that edge will often depend on the sovereignty of their AI systems.
you are also seeing this view expressed in the market with funds such as USV publicly stating they'll be focused on "data at the edge" or @derekedws last two posts touching on the significance of privacy and sovereignty enabled by local inference and open source harnesses, respectively.
the internet freed information, blockchains freed capital, LLMs freed intelligence. new technologies tend to democratize access to previously scarce resources but eventually new forms of scarcity rise to the surface and the moats get rebuilt.
Niftical retweeted
Actual Computer is pleased to announce its first source available-software, toks.
Actual's toks is the best tokenizer on earth and the first tokenizer designed for the quadrillion-token era.
actual.inc/company/blog/intr…
Niftical retweeted
Actual's team thinks about inference every day and lately we've been thinking that
Need is too great to fit inside of just a few buildings
Energy infrastructure is not fit for centralizing power
Bringing us to the long tail of neoclouds
actual.inc/company/blog/the-…
people should have Actual Computers @actualinc
With $zec running i think private AI will become a big narrative the next few months
People should have the right to privately interact with their LLM’s
Sometimes I wonder how people outside of NY are able to survive without their state governments protecting them from dangerous crypto scams such as ETH-backed loans or trading on Kraken
