topmass @topmass

AI Automation Eng • @AnthropicAI Model Safety Tester | Founder https://t.co/bilxNRUQkG | Building in 🇨🇦 northwave.ai British Columbia Joined September 2011

Tweets

23K
Followers

9K
Following

881
Likes

15K

topmass @topmass

20 hours ago

@PeanutTrade @unusual_whales intelligence is there on latest models (pro) but most people don't pay for access to them unfortunately

0 0 2 231 0

View Details

topmass @topmass

a day ago

@_brian_johnson we gonna see I'm gonna keep pushing fast mode unless it runs out super quick

0 0 0 52 0

View Details

topmass @topmass

2 days ago

pour one out for codex fast mode tonight GPT 5.5 party extra usage limits end on June 5th

1 0 0 109 0

View Details

topmass @topmass

2 days ago

literally never using tmux again or going back to anything else herdr is so good

0 0 1 133 2

View Details

topmass @topmass

4 days ago

even codex is sick of waiting around for this dude

0 0 0 203 0

View Details

I used opus 4.8 for a week and making the switch back to codex 5.5 + fast mode (incl in my sub and not EXTRA usage) was the best thing I did. Opus is great when you need have tons of time to wait around for it to finish slow frontend work.

1 0 0 272 0

View Details

topmass @topmass

a week ago

@SteveMoraco it's ok Steve i like you as a chat model

0 0 1 18 0

View Details

topmass @topmass

a week ago

@snapss same it's kind of annoying @nikitabier why algo go brrr canada

0 0 2 83 0

View Details

topmass @topmass

a week ago

The best agent full stack template cloudflare already quietly added about 7 months ago its react router running on a single worker with vite + ShadCN UI and Hono bundled in for the backend models using this are using the stack and UI library they are MOST familiar with

0 0 0 109 0

View Details

topmass @topmass

2 weeks ago

@0xSero @alspinu I hate the gpt ui/ux style on the right - I only use it for backend work since even OSS models do better for me

0 0 0 77 0

View Details

topmass @topmass

2 weeks ago

@0xSero have 5090 and a few other cards with 12-16gb vram - best move stacking them all in a server or just selling and adding another card with more vram ?

1 0 0 355 0

View Details

topmass @topmass

2 weeks ago

@SeedOilDsrspctr tissot PRX under 500 non powermatic but the icy blue powermatic under 900 is a great watch that looks pretty striking

2 0 11 2K 7

View Details

topmass @topmass

2 weeks ago

if you're building ai apps right now the cleanest stack is just react/vite frontend + hono backend on cloudflare workers that's enough reach for react router when seo pages matter, most projects don't need more stack, they need less