@ZssBecker It's because everyone learned not to take you seriously. You were obviously wrong over 95% of the time so just became comedy.
English
Sim
398 posts
@ivanfioravanti We could all be doing this is GPU manufacturers weren't sandbagging consumer gpu ram limits... Tech nazis
English
Kimi K2.6 running locally on 2 x M3 Ultra 512GB! 😱
1T parameters model! 😱😱
Pedro Cuenca@pcuenq
Kimi K2.6 was released 1h ago, and it looks amazing! Here it's running with MLX (mlx-vlm) on two M3 Ultras (full 1T param VLM) 🔥
English
English
What are your initial impressions of Grok 4.20?
Major upgrades are still landing every week.
Testlabor@testerlabor
Grok 4.20 is now officially out of Beta. It's now on Auto, Fast, Expert & Heavy.
English
im fully convinced that LLMs are not an actual net productivity boost (today)
they remove the barrier to get started, but they create increasingly complex software which does not appear to be maintainable
so far, in my situations, they appear to slow down long term velocity
English
@AparryW @BrianMcDonaldIE So peaceful killed tens of thousands of their own citizens for protesting the regime
English
@BrianMcDonaldIE The US sent a war fleet halfway around the world to attack a nation that just wanted peace and sanctions lifted. Nobody came right up to their border and threatened them before they took action. Yet it's vassals around the world will dutifully nod and agree, Iran had it coming.
English
A Russian friend writes on Telegram:
“I have a few questions! Will there be sanctions against Israel and the United States?
Disconnection from SWIFT? Freezing of reserves? Banning of athletes?
Seizure of aircraft? Suspension of air travel?
Or is this different?”
English
@BrianMcDonaldIE Yea it is different... Just look at how Ukraine responded to war coming to them, compared to Iran. It's decided to attack 11 countries. It's not hard to tell who the baddies are.
English
@DerektheCleric @rohitdotmittal Can't vibe code a datacenter with 100,000 gpus
English
@rohitdotmittal The problem is that 5 years people will just vibe code their own Anthropic for free.
English
It's important to note that Anthropic still doesn't have multi-modal capabilities.
They are not working towards AGI.
Their models don't generate images, audio, or video. They are fighting on different fronts than OpenAI, Google, and xAI.
They are eating coding (coders) and enterprise (knowledge work) one vertical, one use case, at a time.
They started with coding, then excel, then banking, then legal, then security, then HR, and more will come soon.
No one is releasing products by vertical at this speed.
They are playing a different game than the rest of the pack.
English
@rohitdotmittal Whichever ai company nails Ai doing ai research.... They win everything
English
@ZssBecker @feifei_qiu They will when u can one prompt it and make it good in the future
Like... Clawbot in two years hey make me a mailchimp clone and deploy it on cheap cloud infrastructure nearby, make sure it's secure, test all works
Agent runs autonomously for a week and makes it Happen
English
Go build a SaaS that manages an important of your business. Like go do it. Build a mailchimp.
Then slowly watch all your time shift from working to managing your mailchimp clone and having to spend hours learning how to keep email delivery up, debugging bugs, staying compliant etc etc and the staff you have to have to manage it.
Then ask yourself ...why would I not just pay mailchimp $250 to do this and forget about it.
Most businesses can't even be bothered to swap out a SaaS product they bought 10 years ago because of the friction it causes.
How many dentist office for example are running on a website made in the early 2010s with a similar call system.
They are not about to start making their own apps and managing them.
English
If anyone can build a SaaS quickly today,
why would someone choose to use yours?
What makes your SaaS defensible?
English
@LewisCTech @LoresongGame Every model they develop is profitable. Ina yearly basis they don't see profitable because the amount they invest into new model training is higher than the revenue from the previous model...but each model itself is still profitable. They are in a race so spending like crazy
English
@LoresongGame WTF are you talking about? Some tech companies in the past became profitable, therefore they all will? That is a very midwit view of the world.
Also if you don't like my posts... mute me? I never reply to yours.
English
I really have nothing against people who prefer vibe coding.
But the bold predictions about the future are a bit much. OpenAI and Anthropic aren't profitable. We really don't know what the future holds.
I'm glad you like but can you all enjoy your vibe coding a bit more calmly?
Ryan Dahl@rough__sea
This has been said a thousand times before, but allow me to add my own voice: the era of humans writing code is over. Disturbing for those of us who identify as SWEs, but no less true. That's not to say SWEs don't have work to do, but writing syntax directly is not it.
English
Figma is very quickly becoming a huge bottleneck in building products.
You build a prototype.
People like the vision and buy into it.
Then Engineers and Product sit and wait till you deliver the whole spec in Figma.
Product & Eng have both become faster due to AI.
Design is getting squeezed in the middle with the old way of working.
English
We shared updates publicly after the 2022 incident, including what happened and what changed.
To clarify, 3Commas does not store exchange credentials or hold user funds. The incident involved a cybercrime where API keys were abused to place trades on exchanges.
Since then, we’ve cooperated with law enforcement, shared identified suspect wallets, and added additional security layers like SignCenter to reduce risk going forward.
English
@simonlukehill We did provide updates after 2022. 3Commas doesn’t custody funds or exchange credentials. The incident was a cybercrime involving abused API keys, and we cooperated with authorities and added new security layers after.
x.com/3commas_io/sta…
3Commas@3Commas
We shared updates publicly after the 2022 incident, including what happened and what changed. To clarify, 3Commas does not store exchange credentials or hold user funds. The incident involved a cybercrime where API keys were abused to place trades on exchanges. Since then, we’ve cooperated with law enforcement, shared identified suspect wallets, and added additional security layers like SignCenter to reduce risk going forward.
English
Spent an hour trying to make AI fix a Supabase auth issue.
Fed it the docs. Explained the context.
Nothing worked.
Every new prompt “you’re absolutely right I should’ve….” then goes ahead to do the same rubbish.
Got mad and swore at it.
Read the docs myself. Fixed it in under 15 minutes.
This is what’s supposed to replace engineers? Come off it.
English
@DanielcHooper Also great at teaching fundamentals as you go. I like using one model for generation, and another for review and reflection.
English
My C programmer thoughts on Claude+Opus 4.5:
1. Bad at writing code: wrote O(n²) algorithm when O(n) possible. I wouldn't commit its code without review.
2. Responds well to feedback "make this algorithm linear". You have to already be a good programmer to know how the code could be improved.
3. Useful for analysis: "How could this system get into ?"
4. Helps get over procrastination on grindy tasks, like creating a linux sysroot for cross compilation.
5. Makes it cheaper to try different approaches: "change the memory layout to X and use data structure Y, run performance test and compare"
6. Running 1 or more agents in the background while I do other work feels like a superpower.
7. Best when treated like a lawyer's paralegal: you do big brain planning, it does tedium in background, you review, tweak, commit.
English
English
@fiat_money @mog_russEN Did NATO provoke this war by orchestrating a coup in 2014 throwing out an elected president and replace that president with one that was pro western then afterwards push for NATO membership? 🤔
English
🚨⚡️ President Vladimir Putin:
"My impression is that NATO needs an external enemy to justify its existence. They are constantly looking for an enemy or provoking someone just to label them as one."
English