Tim Haines 🇺🇦

68.7K posts

Tim Haines 🇺🇦

Tim Haines 🇺🇦

@TimHaines

Prev founded Percy, Respondly, & Favstar. Family guy who enjoys 🍦⛷ Ξ.

Wanaka, New Zealand Katılım Nisan 2008
593 Takip Edilen5.2K Takipçiler
Tim Haines 🇺🇦 retweetledi
Molly Ploofkins
Molly Ploofkins@Mollyploofkins·
Shout out to Qatar for the worst return on a $400M bribe in history.
English
133
4.2K
38.6K
784.3K
Mario Zechner
Mario Zechner@badlogicgames·
why shouldn't i? why shouldn't i be happy?
Mario Zechner tweet media
English
5
1
38
4.4K
Bryan Helmkamp
Bryan Helmkamp@brynary·
AI brings us malleable software. The first beneficiaries should be engineers. Every engineer deserves best-of-breed tools they can evolve when needed. Fabro is open source under the MIT license. Use it at home or at work, with rent and no restrictions. github.com/fabro-sh/fabro
English
1
7
68
4.7K
antirez
antirez@antirez·
Clearly we need some kind of approval process to avoid the ability to open PRs for new accounts without a past GitHub contribution history and so forth. Here the problem is not LLM assisted programming per se, but agents around on the internet doing shit like that.
clem 🤗@ClementDelangue

Our biggest open-source repos are getting overwhelmed by AI slop which literally makes Github unusable (~a new pull request every 3 minutes). Fun new challenges in an agentic world!

English
17
2
69
11.5K
Tim Haines 🇺🇦 retweetledi
Matteo Collina
Matteo Collina@matteocollina·
We benchmarked TanStack Start, React Router, and Next.js running the exact same eCommerce app at 1,000 req/s on AWS EKS. The results were eye-opening.
Matteo Collina tweet media
English
98
209
2.3K
378.6K
Tim Haines 🇺🇦 retweetledi
Joe Kent
Joe Kent@joekent16jan19·
After much reflection, I have decided to resign from my position as Director of the National Counterterrorism Center, effective today. I cannot in good conscience support the ongoing war in Iran. Iran posed no imminent threat to our nation, and it is clear that we started this war due to pressure from Israel and its powerful American lobby. It has been an honor serving under @POTUS and @DNIGabbard and leading the professionals at NCTC. May God bless America.
Joe Kent tweet media
English
72.4K
219.6K
845.6K
98.9M
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@thsottiaux Permissions. Let me permit read operations and get asked for write permissions. Allowlisting would be 💯
English
0
0
0
28
Tibo
Tibo@thsottiaux·
What are we consistently getting wrong with codex that you wish we would improve / fix?
English
1.2K
14
872
141.5K
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@robdel12 @Legendaryy I’d guess average utilization is well below 50% too - with most people getting no-where near the max use out of their plans.
English
0
0
1
24
Rob the builder
Rob the builder@robdel12·
@Legendaryy This is all based on the prices that Anthropic sets. No one but them knows the actual inference costs and I’m willing to wager they’re not lighting money on fire. I bet those margins are cushy.
English
1
0
3
231
Legendary
Legendary@Legendaryy·
Your Claude subscription is massively subsidized and it won't last forever. A $20/mo Pro plan burns through ~$180/mo in API-equivalent tokens. Heavy Max users hit $5K/mo on a $200 plan. Actual compute cost is roughly 10% of API pricing. Venture capital covers the rest. Anthropic just 2x'd Claude Code limits for "Spring Break'" during off hours. Enjoy it while it's here. At some point the math has to math. Clip from the @modernmarket_
English
237
115
1.8K
333.4K
Tim Haines 🇺🇦 retweetledi
Rob the builder
Rob the builder@robdel12·
What if you had @OpenAIDevs's symphony but: - A rich native UI to manage / see the agents - Interact with the agents as they run - Claude & Codex - Any pluggable trigger (Linear, GH issues, WS) - Rust server that can be deployed to any container github.com/Robdel12/Orbit…
English
1
1
0
165
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@bentlegen What was your prompt? You didn’t tell it to maintain the number of tests or keep the test integrity intact?
English
1
0
0
280
Ben Vinegar
Ben Vinegar@bentlegen·
Useful plugin but you have to be careful what you specify. e.g. I asked it to optimize our test suite wall clock time. After 60 iterations, Codex 5.3 split the suite in 2 parts – fast and slow – then only ran the fast part and declared success.
tobi lutke@tobi

OK, well. I ran /autoresearch on the the liquid codebase. 53% faster combined parse+render time, 61% fewer object allocations. This is probably somewhat overfit, but there are absolutely amazing ideas in this.

English
10
0
46
13.9K
Tim Haines 🇺🇦 retweetledi
Kelly Eckhold
Kelly Eckhold@kellyenz·
Here's an overview of the sorts of measures that other countries in the Asia Pacific region are taking in response to the Iran war.
Kelly Eckhold tweet media
English
22
335
777
70.3K
Rob the builder
Rob the builder@robdel12·
Frick, I’m really going to have to touch grass
Rob the builder tweet media
English
1
0
2
64
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@0xTab Ahh! This would probably be a good add to the docs - unless I missed it in there somewhere.
English
1
0
0
20
Tabish
Tabish@0xTab·
@TimHaines Hmm, in theory yes but usually I’m a bit more cautious and for my sanity I would explicitly say something like: /opsx:apply we’ve updated the plan and changed the tasks, find the best way to continue
English
1
0
1
16
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@0xTab what's the best video to watch to get up to speed on the latest openspec, starting from never using openspec before?
English
1
0
0
136
Tim Haines 🇺🇦
Tim Haines 🇺🇦@TimHaines·
@0xTab Your docs are pretty decent. If you start apply, you can pause, edit the artifacts, and apply again, and it’ll do something sensible?
English
1
0
0
30
Tabish
Tabish@0xTab·
@TimHaines but honestly the best way to start is to just start using it IMO and learn along the way :))
English
1
0
0
58
Daniel Sempere Pico
Daniel Sempere Pico@dansemperepico·
You guys all run Claude Code with claude --dangerously-skip-permissions right? Because otherwise how in the world can you sit there accepting every single permission when building something?
English
474
22
2.2K
285.5K