Setting up camp at 👀 engramsoftware.bsky.social

While I did help a little startup find the right developers for their project, I will never, ever be posting a small job on LinkedIn ever again. RIP my email and the contact form of my website (that is not what it! is! for! [internal screaming])

Excited to finally be giving this talk at Microsoft tonight! If you missed the sign up or are otherwise unable to make, I think it may end up being recorded. If it's not, I'll record a version and put it up anyways.

Thrilled to introduce RefuelLLM-2, our latest family of LLMs built for data labeling and enrichment tasks. RefuelLLM-2 (83.82%) outperforms GPT-4-Turbo (80.88%), Claude-3-Opus (79.19%), Llama3-70B (78.2%) and Gemini-1.5-Pro (74.59%) on a benchmark of ~30 data labeling tasks: RefuelLLM-2-small (79.67%), aka Llama-3-Refueled, outperforms all comparable LLMs including Claude3-Sonnet (70.99%), Haiku (69.23%) and GPT-3.5-Turbo (68.13%). We’re open sourcing the model: huggingface.co/refuelai/Llam… You can try out the models here and give us some feedback! labs.refuel.ai/playground. The code and data used for benchmarking the LLMs is available in our Autolabel library: github.com/refuel-ai/autolab… One more thing: RefuelLLM-2 family of models output much better calibrated confidence scores - a useful lever to reject, retry or ensemble low confidence outputs.

When you find ListViews with a hard coded count of 1, you know you're in for some shit 😳

If you're in NYC, I'll be giving a talk on LLM systems at the Microsoft offices on May 13th - it's free to attend, and I'd be really happy to see you there meetu.ps/c/55LvF/lSsBT/a

Working on reinventing the browser in the context of the latest AI boom, I feel like I did in 2012 when I was charting unexplored web APIs like shadow DOM. Uncharted territory, no map. Creating the map is the work.

Kind of coming full circle over here, I'm working on a web browser again. Think "some of that AI stuff, in the browser, but not annoying and doesn't suck" My goal with any software I write is "not annoying and doesn't suck" and anything else is gravy 😅

I still think it's a missed opportunity that @Microsoft didn't bring back Clippy when they got on the AI-assistant bandwagon

One crucial skill about leadership positions is to be able to internalize that the life is not a zero sum game. Anyone who fail in this basic recognition eventually fail because they are swimming against the current.

Tired of finding “this article was generated by AI and may not be accurate” on programming tutorial websites 😔

The senior managers I know going back to IC roles are able to execute like senior ICs and still able to code. Their understanding of organizational nuances also gives them super powers in a Staff role. So even if they are a little rusty on modern tech, it makes up for it.

I never thought "one day I'll be a manager." I was in a situation, twice, where there was a need for a manager. In one case, I just did it. In the other, we tried to hire for 6 months, but couldn't. So I took it on. It was situational. Required reading: charity.wtf/2017/05/11/the-e…

Another engineering manager I know changed jobs to become an engineer again, and it makes me happy to see that the manager - engineer "pendulum" (as @mipsytipsy calls it) is a lot more common these days. Being a manager is a role: and engineers sometimes take it up. For a while.

I wish that it was an inherit requirement from the engineering management ladder to be able to switch between IC and management roles. Mid level managers end up executing the job as TPMs when they are disconnected from engineering only after a couple of years.