490 karma · joined May 3, 2017
if you already have claude, you can even get it to set it all up for you too
the issue with local models is just the compute required, it's hard to compete with 100t/s cloud models
that being said, deepseek v4 needs to be on amazon bedrock to actually be feasible in the US Enterprise market and start driving other provider prices down
- Broad slice:
- Full Forge: 48/72 accurate, 72/72 complete, score 66.7%
- Bare: 18/72 accurate, 24/72 complete, score 25.0%
- Lift: +30 correct runs, no paired regressions
- Bare had 42 ToolCallErrors and 6 ToolExecutionErrors; full Forge had none.
- Advanced reasoning:
- Full Forge: 3/24 accurate, 24/24 complete, score 12.5%
- Bare: 3/24 accurate, 9/24 complete, score 12.5%
- Lift: completion improved, but accuracy did not.being able to program is not the only skill required to be a successful software engineer, so no ai agents cannot be software engineers
very important distinction - i personally like the radiologist example - looking at scans is a part of a radiologist job, AI can do it better than most of them, but looking at scans is a small part of the job, most of it communicating with doctors to help their patients
i can kinda see it, they spent a lot of time getting Gemma 4 pretty efficient and then seeing everyone buy macs to run them and realize it’s maybe a real moat since Apple doesn’t make any AI
Would be an interesting product if it could actually give you GPT performance locally, will be an awful experience if it’s essentially just cloud AI…like a premium laptop where most of the features are locked behind a subscription would be wild
I don't think agentic workflows are there yet, but implementing skills to manually call and use while working side by side with an AI is definitely nice - our company is focused a lot on sandboxing right now and having safe skills
I don't think we've gotten feature development well yet, but the review skills + grafana skills they wrote have been pretty solid
with superpowers, i see a lot of specs -> impl plan -> execute plan
Would definitely love benchmarks against omlx and fast-mlx one day (i also have a 256gb m3 ultra)
knowing when and when not to use k8s, is also a skill
It's such a waste of time and resources - you wanna handle gun violence? handle normal violence with proven mechanisms (education, social welfare, etc...)