@thsottiaux Its unusable the past 12 hrs , it would run for 30mins and output rubbish then revert.
English
Sean Turing
1.2K posts
@MiniMax_AI Minimax made me feel like a developer again because i had to provide detailed prompts since pricing is based on per prompt 😂
English
Sean Turing retweetledi
M2.5 has been live for 3 days — would love to hear your thoughts!
To enter:
1️⃣ Repost this post
2️⃣ Leave a comment
We’ll pick the Top 5 most liked comments to win 3 months of Coding Plan HighSpeed Plus ($120 value) 🎁
You can share any thoughts, opinions, or stories about your experience with the model — we’d love to hear them.
platform.minimax.io/subscribe/codi…
MiniMax (official)@MiniMax_AI
MiniMax M2.5-HighSpeed⚡is live! 100 TPS — enjoy the 3× speed. In the 48 hours since launch, thank you all for your incredible support and love for MiniMax M2.5! Designed for the next generation of Agent applications, we’ve officially launched MiniMax-M2.5-HighSpeed. Delivering blazing-fast inference at 100 TPS — 3× faster than similar models. Supports both API integration and Coding workflows. Choose from Plus, Max, or Ultra plans. Invite friends and enjoy 10% off instantly. Wishing you all Happy Chinese New Year! May the 🐎 year bring speed, scale, and breakthrough luck. 🚀
English
GLM-5-Code will probably drop soon , sounds like a infrastructure issue. Imagine if china could buy GPUs
English
Every frontier AI lab has lost co-founder(s):
- xAI: 5 of 12 gone, 1 seriously sick
- OpenAI: 8 of 11 gone
- Thinking Machines: 3 of 6 gone
- SSI: 1 of 3 gone
- DeepMind: 1 of 3 gone
Except Anthropic. All 7 co-founders are still there.
What’s Anthropic’s secret?
English
bruh
opus 4.6 fast is SIX TIMES more expensive and ONLY 2.5x faster
who is this even for?
English
@tombielecki @amasad Qwen feels like bootstrap now , never existed
English
Chinese new year is in less than 10 days. This is probably DeepSeek 🙂
OpenRouter@OpenRouter
🥷 We’re launching a new stealth model on OpenRouter: Pony Alpha. - Pony Alpha is a next-generation foundation model - It delivers strong performance across coding, reasoning, and roleplay - It’s optimized for agentic workflows, with high tool-calling accuracy
English
English
Sean Turing retweetledi
claude-sonnet-5-20260204
It's here, yesterday was false positive 😗
English
4,073,000+ companies and 2.8m founders and execs organized into targetable lead lists.
No scraping. No cleaning. Just plug and play.
I took 2 massive company database and segmented it into 13 industry-specific lists.
Here's what's inside:
→ 817K Software & Technology companies
→ 746K Diversified/Other industries
→ 551K Manufacturing & Industrial
→ 352K Healthcare & Life Sciences
→ 294K Media & Entertainment
→ 292K Real Estate companies
→ 270K Financial Services
→ 189K Professional Services
→ 189K Retail & E-commerce
→ 128K Education companies
→ 101K Food & Beverage
→ 75K Transportation & Logistics
→ 64K Energy & Utilities
BONUS data included:
→ 2.8M Executives & Founders (C-suite, directors)
→ 765K Funding Rounds (seed to IPO)
→ 582K Investors (VCs, PEs, angels)
→ 6,500 Strategy Consulting firms (McKinsey-types)
8.5M+ total records across 151 files.
Each list includes: Company name, description, website, location, employee count, LinkedIn, industry, ownership status, funding data.
Perfect for cold outreach, market research, or building your ICP.
Like + Comment "LISTS" and I'll send you the drive link.
Gotta be following to receive DM.
English
I've came to realise that if you face issues with @Cloudflare as a paying customer even with proper process, it will be stuck in "In Review" or support for more than 1 month.
Someone should build an alternative , this monopoly isnt good.
English
Oh wow that list of contact information , great for sales
Jonathan T.@JohnDotOwl
Is this good enough for release? I built this to discover niche mobile apps that I can build some but i feel it's only 50% done. Let me know if you want to try it out.
English
@htihle It’s easy to spot which benchmarks are not accurate right now because I’m actively using Gemini flash 3.0 beside opus , flash is a really good model
English
Gemini 3 flash scores 61.6% on WeirdML, beating out gpt-5.1 and pushing the frontier of cost vs accuracy.
For reference gemini-2.5 flash and pro scored 41.9 and 54.0%, so this is quite a step up for flash.
One thing that stands out for this model are the code execution times bunching up close to the maximum two minutes. This is the first model to so consistently use all the available resources.
Håvard Ihle@htihle
WeirdML v2 is now out! The update includes a bunch of new tasks (now 19 tasks total, up from 6), and results from all the latest models. We now also track api costs and other metadata which give more insight into the different models. The new results are shown in these two figures. The first one shows an overview of the overall results as well as the results on individual tasks, in addition to various metadata. The second figure shows cost vs performance and shows a clear scaling with better results for higher costs. We also have a very varied pareto frontier with 11 models from 6 different companies having the best accuracy for a given cost for at least some of the cost range. Grok 3, Claude Opus 4 and GPT 4.5 are the ones that underperform for their costs, while Gemini pro and o3 pro have the best results at the highest costs. Qwen3 30B3A, grok 3 mini and deepseek R1 also each represent a good chunk of the pareto frontier.
English
in singapore for the next 3 days.. what should i check out?
English
@arpit_bhayani GPT 5.2 does it faster , anthropic model does this too but it’s usually a mistake or they don’t even notice. And more often than not it’s related to migration , rollback , seed commands
English
I was vibecoding and asked Claude to add a feature, and it legit (no kidding) cleaned up my database. I always thought such incidents were made up, but now that it has happened to me, I get it.
English
@TosinOlugbenga Oh wow I better be writing my own pagination code from scratch then , don’t use any libraries
English
A software engineer that can’t implement its own auth….
Is that one a software engineer?
The least basic thing in the whole software architecture is authentication…RBACs…
I wonder why some people will think running Auth as a service is a thing…
If I am the CTO of a company and any engineer suggest we subscribe to a third party auth system, I will fire the person.
chinedu🦀@chinedu_10
Every engineer should be able to implement their own authentication. The fact that this is even a conversation is scary. What do you mean you can’t implement your own auth and must use an external or third-party service because some tech influencer told you it’s best practice? Are you stooopid or something?
English
@TheGeorgePu I mean because you kinda want them to be your employee so you have better returns long term , let’s face it
English
Rejected a founder yesterday.
$320K ARR. Growing. Profitable.
First question he asked: 'How fast can you get me to $10M?'
Me: 'Why $10M?'
Him: 'That's the exit number I need to sell it for 5x.'
Building for an exit you don't have is building for no one.
I only work with founders who want to own it forever.
English
@TheGeorgePu Is it safe to say that it’s a pretty good idea to compete against products with VC funding because they will probably burn out
English
Know a founder who raised $8M in 2021.
Valuation: $40M
His equity: 30%
Paper worth: $12M
Today:
- Can't raise Series A
- Burning $150K/month
- 4 months of runway
- Interviewing at Apple
The equity? $0.
He spent 4 years building someone else's dream on his calendar.
That's the VC trap.
English
The new 0.7 release of @kirodotdev came out public yesterday, so here are some of my thoughts on the latest features:
- Kiro Powers
- Autonomous Agent
- Context usage meter (finally!!!)
- Context summarization
- Slash commands
Open the thread to read more.
English
@Liberty_cap_M @nextjs LOL no wonder so many servers are under attack
English
@nextjs Wtf is happening? I also recieved a email from vercel.. is it really important to update next version in my saas plaform?
English
Update on CVE-2025-66478 (React2Shell):
An npm package has been released to scan and update affected Next.js apps. Use `npx fix-react2shell-next` to update to patched versions.
All users should update as soon as possible.
More details our blog:
nextjs.org/cve-2025-66478
English