@metaskills

Innovative & Award-Winning Leader | AWS Hero | AI Business Impact thru Strategic Tech, Product, Marketing, & Sales. @torcdotdev

 Portsmouth, VA
unremarkable.ai
Joined November 2007
  • Tweets41.1K
  • Following1.2K
  • Followers2.6K
5.8KMedia
🚀 The AI Speed vs Intelligence Trade-off: A Critical Inflection Point unremarkable.ai/exploring-ac… Fascinating developments in AI this month: While OpenAI's o3 and Google's Gemini 2.0 are pushing the boundaries of reasoning capabilities, they come with a shocking cost - o3 took 5-6 minutes PER TASK and cost $350K for the ARC AGI benchmark run! But here's the game-changer: Working with @SambaNovaAI's Cloud platform, we've achieved 4x faster inference speeds with Llama 3.3-70B compared to traditional platforms. This means we can build Compound AI systems (multiple specialized models working together) that are both intelligent AND responsive.

Exploring Accelerated Compound AI Systems with SambaNova & Llama 3.3-70B

OpenAI o1 is approximately 30 times slower than GPT-4o. Similarly, the o1 mini version is around 16 times slower than GPT-4o mini. Frontier models are getting slower! Yup, you heard that right. A few...

unremarkable.ai
1
1
6
28,532
this mf'ers would get excited about a Briggs & Straton DIY hovercraft from the back of an 80s popular mechanic if elmo released it
Sawyer Merritt
@SawyerMerrittOct 17
These are the Tesla Roadster’s deployable skirts, which work with four rear-mounted fans to generate enough suction to hold the vehicle upside down, even when completely stationary. It'll help with track performance too.
3
243
blindly blaming the others at all costs
DHH
@dhhOct 18
Mass immigration has been a catastrophe for Europe. There is no issue more urgent before the continent than to correct this mistake.
2
334
THIS! Also don't y'all know we have had the technology to make a fucking brick do this for like what... decades. Kick flips in cars is just marketing.
Corey Quinn
@QuinnyPigOct 16
This is absolutely someone at Tesla shaking the keys to distract Elon from fucking up something important.
163
🚨 Mods aren’t sandboxed❗️ This opens the door for Skills in Desktop that rely on general computer use. That's a game changer! code.claude.com/docs/en/plug…

Mods overview - Claude Code Docs

Add panes, commands, and tool call rules to Claude Code with a mod. See what a mod can do, how to make or install one, and where mods run.

code.claude.com
104
Thanks @__steele for the amazing write up on the new Lambda Web Functions. awsteele.com/blog/2026/10/12…

Some notes on Lambda Web Functions

Some notes on Lambda Web Functions

awsteele.com
1
2
165
Ken Collins@metaskillsOct 14
ordered
131
ordered
1
1
220
I hate I want one.
Evan Kirstel #B2B #TechFluencer
@EvanKirstelOct 13
Replying to @johnternus
@johnternus
162
The Node.js web app on your laptop can now go on Lambda and keep listening on a port, with no adapter in between. Lambda Web Functions launched yesterday, and one copy of your server takes up to 64 requests at once! @awsdevelopers #Lambda

AWS launches Lambda Web Functions: run your Express app on Lambda

The one line you change, and why a module-level currentUser passes every test on your laptop but leaks one user's data to another on Lambda.

1
7
1
44
1,475
Been doing stateless MCP on Lambda and this makes me want to use it more.
Matt Carey
@mattzcareyOct 12
Replying to @danieljvdm @thomas_ankcorn
It always could. They were called notifications. When the MCP spec when stateless it got a bit annoying tho to have a persistent connection just for this. The new MCP Events/triggers spec allows for webhooks to deliver events :)
1
2
257
about a year ago it was clear to me that writing software and delivering products would never be the same, but #RSI takes it to a new level, i'm blown away
1
3
302
Ken Collins: the agent harness is heading below the waterline. Important, but nobody will care about it. Chris Farris: the model is swappable. The harness is what you build your workflows around. Yells at Cloud Ep. 3: youtu.be/ws83xEExg_Y
1
47
We opened Yells at Cloud ep. 3 by calling AI "super intelligence" as a joke. On Tuesday it became an executive order. Agencies now have to say "Super Intelligence" and "SI" instead of "AI." yellsatcloudpod.com/episodes…
1
1
2
47
Dweller
@One_Way_HomeApr 6
Not a huge fan of SNL, but I gotta admit, this had me rolling…
1
110
lol to anyone peddling Openclaw. there is value out there, it is a big distraction and not helping.
⭕ Brock Pierson
@brockpiersonMar 7
Openclaw is one big grift. Nobody is building anything real. It's just grifting influencers telling YOU how to build stuff (but never build anything themselves). Interesting how you can be building 24/7 but have nothing to show for it... Check the track record of these snake oil salesmen. Same type of person who polluted crypto with their grifty ways. Selling a false dream and taking money from innocent people is disgusting. You can't change my mind.
2
1
4
1,012
Every Mac can setup Command+Command voice dictation. Use your platform's built in dictation.
Thariq
@trq212Mar 3
Voice mode is rolling out now in Claude Code. It’s live for ~5% of users today, and will be ramping through the coming weeks. You'll see a note on the welcome screen once you have access. /voice to toggle it on!
2
275
Just in time!
Christina Warren
@film_girlFeb 13
Happy Galentine's Day! You can now disable PRs on GitHub or limit them to contributors to your project! Amazing work from @helenhousandi, Camilla Moraes, Clay Miller, Stephanie Hang, Jan Maarten and more!
177
Opus 4.6 is killing me! I might have to switch back to 4.5.
182
We are in the managed runtime loop wars now.
ᴅᴀɴɪᴇʟ ᴍɪᴇssʟᴇʀ 🛡️
@DanielMiesslerJan 26
This is because it's no longer about models. GPT 5.3 could be 10%, 20%, or even 50% better than Opus 4.5 but it wouldn't matter. Scaffolding is what matters. And OpenAI is many generations behind on that.
117
claude still thinks it is 2025
1
78
Load more