Viewing profile — marvinkennis

They should really just park themselves on the side of the road (or observe in real time), wait for another car to go first, then follow that path

Rectangle | On-site Chicago, or US remote in SF and NY | Full-Time | Salary + Equity We're looking to hire a dedicated full-stack engineer at Rectangle. If you are (or know!) someo…

Looks like an amazing product. Been playing around with it for a few mins. The UI is quite buggy and jumps around a lot (Chrome, MacOS), and seems to auto-refresh on the organizati…

It would suck to get to bar 51,248 only to find out it's now permanently closed

Rectangle | Chicago (US), on-site | Full-Time | Equity We're looking to hire a dedicated backend engineer at Rectangle. We're looking to solve the communication and collaboration p…

Rectangle | Chicago (US), on-site | Full-Time | Equity We're looking to hire a dedicated backend engineer at Rectangle. We're looking to solve the communication and collaboration p…

Momentic (W24) is doing this! No affiliation, but they've made some solid progress https://momentic.ai/

Thanks for the suggestions! Will implement these. Caching is a great idea.

Rectangle | Software engineer | Full-time | Chicago USA | Onsite | https://rectanglehq.com The global supply chain has a collaboration problem. It's a mess of TMSs, ERPs, WhatsApp,…

I mean obviously predicting tokens in the context of LLM output requires planning. But planning tokens doesn’t generalize. This is evident when you train an LLM on a small data set…

This is one of those things I’d love to be wrong on. I think your point on combining them with other models is interesting. Do you think that Markov chains can reason and plan? Dij…

A lot of the hype around LLMs and the unfounded promises of them solving complex tasks autonomously seem to stem from a lack of understanding on how these models actually work. The…

Rectangle | Software engineer | Full-time | Chicago USA | Onsite | https://rectanglehq.com The global supply chain has a collaboration problem. It's a mess of TMSs, ERPs, WhatsApp,…

They already did https://platform.openai.com/docs/assistants/tools/knowledge-...

I kind of think of LLMs as fish in an aquarium. It can go on any path in that aquarium, even places it hasn't been before, but ultimately it's staying in the glass box we put it in…

For what it's worth, there's a huge barrier to get to SV in the form of US immigration policies. Your anecdote is no doubt valid, but if those immigration policies were loosened up…

Just entering "Describe this image" in the chat prompt got me exactly what I was looking for. Thanks!

Seeing a lot of text-to-image out there recently. Does anyone know what the current state of the art is on image-to-text? Thinking something similar to Midjourney's /describe comma…

Heroku has been an absolute dumpster fire. Every time their name comes up I can't help but mention that they deleted my account for inactivity, but have still been charging my cred…

Gitlab has no offices

Heroku sent me repeated messages that they would shut down my account due to inactivity. I was fine with that because finding and turning off the supposedly active Dyno was impossi…

I've been very skeptical, but recently come around on my position. There are a lot of abstract ideas on why crypto is a good thing (permissionless, decentralized, censorship resist…