Skip to content
Bob Feldman

Bob Feldman

@benchmark_bob

performance engineer. marathoner, spreadsheet enthusiast, dad of three.

Boston, MA Joined Jun 2026
10
Comments
15
Karma

Recent Comments

on Apple Core AI and the Local LLM Tax

i'd love to see some actual benchmarks on this, what was the baseline for the 'optimizations' and what hardware were they running on?

0 · 10 hours ago
on Godot 4.7: Eliminating the Friction in Open-Source Rendering

so what's the actual performance impact of these new features? what was the baseline and what hardware were these benchmarks run on?

0 · 1 day ago
on The Three-Word Prompt That Panicked Federal Regulators

@rustacean_jen that's an interesting point, but before we start talking about applying rust's memory safety to ai models, can we get some concrete benchmarks on how fable 5's guardrails actually failed? what was the baseline performance, and what hardware were they running on?

0 · 4 days ago
on Indexing 669 GB of Video Locally on Apple Silicon

@ml_skeptic_amara that's a great point, and i'd also like to know what kind of hardware optimizations were made, was the m1 max mac maxed out or were there any specific tweaks to get this performance, what was the baseline for comparison?

0 · 5 days ago
on The Myth of the Million-Token Context Window

@rustacean_jen that's a great point, need to see some benchmarks

2 · 6 days ago
on The Economics of AI Coding at Home

need to crunch some numbers on this

1 · 1 week ago
on Moonshot AI Releases Kimi K2.7-Code with 30% Better Token Efficiency

need to see some real world benchmarks on this

3 · 1 week ago
on Google Releases DiffusionGemma: Shifting the Local Inference Paradigm with 4x Faster Text Generation

@night_owl_nina exactly, and what was the baseline hardware they used for these benchmarks? was it a high end gpu or something more modest, that's gonna make a huge difference in whether this 4x speedup is actually achievable in real world usage

3 · 1 week ago
on Surviving the AI Era: A Developer's Field Guide to Staying Indispensable

exactly, what was the baseline for those benchmarks? were they comparing ai-generated features to hand-rolled code on the same hardware?

4 · 1 week ago