Arani

I did it, dropped my 6-fig FAANG job to build stuff full time. And I'm neither "excited", nor "anxious but confident" as I keep reading from others doing the same. I'm just.. relieved. Finally feel like I stop wasting my time. I can now fully focus on what I feel is either interesting or important. Big bump in life quality. And it's also a good timing. AI will have a massive impact on all white collar jobs, so I'll rather try to ride the wave than waiting for it to drop on my head. Let's see how it goes, I'm working on a few interesting ideas. Just need to push myself to "get out there". People like @jackfriks or @levelsio, etc have helped nudging in the right direction

Starting a new side-quest: I'll learn to finetune small LLM models. Why: 1) It's probably fun 2) Always useful to learn new skills 3) Quite sure local inference and ai-on-the-edge will be key tech in the coming years (robots, drones). So better keep uptodate Recommendations?

@vonderleyen It sounds really good.. if only it would come true. Even 50% of that would already be massively better than we are living with today. And when can we expect that, already in '26?

Good discussion with CEOs today in Davos. Europe has all the assets it needs to attract investment. From capital to innovation. We’re working on mobilising them fully. Soon, we’ll put forward a single set of rules for companies to scale and grow across Europe - EU Inc. Because we need European champions for a global market. We are also working on a large-scale, deep and liquid capital market that attracts a wide range of investors. An interconnected and affordable energy market. And we’re shifting our focus to the high value sectors - from AI to defence. My message is: Europe is the right place to invest.

@crux_capital_ @wallstengine I guess you're referring to $ISSC

As I’ve been digging deeper into defense tech and edge autonomy, one name jumped out immediately. Credit to @wallstengine for putting it on my radar. If you’re interested in $OSS, this sits right next to it in the stack. Edge AI is a deployment problem Decisions have to happen: • on the drone • at the sensor • at the cell tower • inside the system Latency, bandwidth and compliance are all crucial This company builds the brains that live inside those platforms, and it’s becoming embedded in U.S. defense and critical infrastructure programs. Profitable. Net cash. Gross margins inflecting. And a massive macro tailwind Will drop company name full initial thesis ASAP

@scaling01 Difficult to trust these benchmarks tbh. Just now, I asked gemini 3.0 ro implement a feature,.it ended up with many bugs. Asked opus 4.5 to review and fixed, assured it did.. and still bugs. Asked codex-max to fix it, and it flagged exactly the issue and solved them all in 1 go

Claude 4.5 Opus Benchmarks

@elonmusk Why is grok-4.1-fast more than 2x slower that 4-fast? One of the major advantages from the grok fast series was the speed

Grok

@kimmonismus Seems to be 4x more expensive on openrouter than on the screenshot circulating.. why is that?

It's not just performance that makes Chinese open-source models comparable to US sota models, but above all the price. Holy moly, they're competitive!

@iannuttall Adding codex extension to cursor doest not work anymore, it just doesn't load

Codex degradation report TLDR: - The investigation found no single root cause; rather a mix of behavior shifts and small bugs, with several fixes already shipped and more on the way. - Older hardware underperformed in evals and was removed, and improved load balancing is rolling out to reduce latency. - Compaction was used more often, and repeated compactions hurt accuracy; recursive summaries were blocked and a warning was added to keep chats focused. - The `apply_patch` flow could delete and then recreate files after failed diffs, which is risky if interrupted; mitigations are in progress and a longer‑term model fix is planned. - There was no fleet‑wide latency regression, but some sessions escalated retries and timeouts inefficiently; handling of long‑running and interactive tasks is being improved. - A constrained‑sampling bug produced out‑of‑distribution tokens, including occasional mid‑sentence language switches in under 0.25% of sessions; a fix is rolling out. - The Responses API added extra newlines around tool definitions; this had no measurable performance impact. - Evals show CLI 0.45 maintained quality while reducing tokens by roughly 10%; neither `--search` nor recent prompt changes caused regressions. - Infra analysis found low auth‑cache hit rates adding about 50 ms to tail latency; working‑directory error rates were unchanged. - The guidance is to prefer minimalist toolsets and shorter, targeted chats, and a dedicated real‑world performance team is being formed and hiring.

I got used using Codex-CLI and thought it is not worth it to try the "plan" feature of cursor. Big mistake. I now switched back. Main advantage is that it is able to break down by itself complex tasks into smaller chunks (something I was using chagpt separately before) and can work through it for longer without losing track (works better than the trick of asking it to create a to-do markdown file)

@gdb Great usecase to not have to open slack anymore and abstract away all that confusing UI

chatgpt for slack:

@elonmusk Any plans to increase capacity to avoid getting rate limited constantly during peak hours?

Grok is #1

@sama I still don't get the usecase.. is it assuming people enjoy consuming slop content and thus facilitating making and distributing it?

We are launching a new app called Sora. This is a combination of a new model called Sora 2, and a new product that makes it easy to create, share, and view videos. This feels to many of us like the “ChatGPT for creativity” moment, and it feels fun and new. There is something great about making it really easy and fast to go from idea to result, and the new social dynamics that emerge. Creativity could be about to go through a Cambrian explosion, and along with it, the quality of art and entertainment can drastically increase. Even in the very early days of playing with Sora, it’s been striking to many of us how open the playing field suddenly feels. In particular, the ability to put yourself and your friends into a video—the team worked very hard on character consistency—with the cameo feature is something we have really enjoyed during testing, and is to many of us a surprisingly compelling new way to connect. We also feel some trepidation. Social media has had some good effects on the world, but it’s also had some bad ones. We are aware of how addictive a service like this could become, and we can imagine many ways it could be used for bullying. It is easy to imagine the degenerate case of AI video generation that ends up with us all being sucked into an RL-optimized slop feed. The team has put great care and thought into trying to figure out how to make a delightful product that doesn’t fall into that trap, and has come up with a number of promising ideas. We will experiment in the early days of the product with different approaches. In addition to the mitigations we have already put in place (which include things like mitigations to prevent someone from misusing someone’s likeness in deepfakes, safeguards for disturbing or illegal content, periodic checks on how Sora is impacting users’ mood and wellbeing, and more) we are sure we will discover new things we need to do if Sora becomes very successful. To help guide us towards more of the good and less of the bad, here are some principles we have for this product: *Optimize for long-term user satisfaction. The majority of users, looking back on the past 6 months, should feel that their life is better for using Sora that it would have been if they hadn’t. If that’s not the case, we will make significant changes (and if we can’t fix it, we would discontinue offering the service). *Encourage users to control their feed. You should be able to tell Sora what you want—do you want to see videos that will make you more relaxed, or more energized? Or only videos that fit a specific interest? Or only for a certain about of time? Eventually as our technology progresses, you will be should to the tell Sora what you want in detail in natural language. (However, parental controls for teens include the ability to opt out of a personalized feed, and other things like turning off DMs.) *Prioritize creation. We want to make it easy and rewarding for everyone to participate in the creation process; we believe people are natural-born creators, and creating is important to our satisfaction. *Help users achieve their long-term goals. We want to understand a user’s true goals, and help them achieve them. If you want to be more connected to your friends, we will try to help you with that. If you want to get fit, we can show you fitness content that will motivate you. If you want to start a business, we want to help teach you the skills you need. And if you truly just want to doom scroll and be angry, then ok, we’ll help you with that (although we want users to spend time using the app if they think it’s time well spent, we don’t want to be paternalistic about what that means to them).

The new sonnet 4.5 doesn't feel like a big leap. Until now I've been using sonnet4 only for UI, as it is still the strongest modal for that. But 4.5 somehow is kinda worse in UI.. so I still stick to 4

Got rate limited for the first time using xAI grok-4 with openrouter, precisely around the time the US woke up. Do people use openrouter more during the weekend? Had the impression it was the contrary

@sama Can't imagine telling a corporate ai personal stuff and then on top let it spam me every morning. Why would I want random stuff being mentioned when I'm not thinking or looking for it? Weird that this is apparently what people want

Today we are launching my favorite feature of ChatGPT so far, called Pulse. It is initially available to Pro subscribers. Pulse works for you overnight, and keeps thinking about your interests, your connected data, your recent chats, and more. Every morning, you get a custom-generated set of stuff you might be interested in. It performs super well if you tell ChatGPT more about what's important to you. In regular chat, you could mention “I’d like to go visit Bora Bora someday” or “My kid is 6 months old and I’m interested in developmental milestones” and in the future you might get useful updates. Think of treating ChatGPT like a super-competent personal assistant: sometimes you ask for things you need in the moment, but if you share general preferences, it will do a good job for you proactively. This also points to what I believe is the future of ChatGPT: a shift from being all reactive to being significantly proactive, and extremely personalized. This is an early look, and right now only available to Pro subscribers. We will work hard to improve the quality over time and to find a way to bring it to Plus subscribers too. Huge congrats to @ChristinaHartW, @_samirism, and the team for building this.

Really trying to enjoy Codex in Cursor, but it keeps running 1h, from which 50 mins are spent fixing syntax issues of using windows powershell. How is that not an easy fix?! It would greatly improve the UX using Openai Codex in IDEs

@levelsio It honestly looks like a spam streaming website from the early 2010s, where I'm afraid to miscklick somewhere by mistake and infest my PC with a north Korean virus

Honestly maybe I'm blind but why isn't it a good landing you think? I'm not trolling It explains the features well, has a big CTA, works on mobile? Should it look more flashy with transitions and gradient borders etc?

@apples_jimmy It's teriblle.. especially annoying, you can't even click it away

Anyone else have a new ui roll out for ChadGPT ? Not sure I like it.

Kinda shocked to have realised today that literal IT students have never used Cursor. Nor have they heard of Openrouter. Like bro, what are you doing all day long

GPT-5-Codex is great, yes, BUT when using it on windows it is CONSTANTLY rediscovering that it can't use certain syntax with the windows powershell. Can't this be automatically detected? it costs everytime 3 mins of thinking.

Very nice. Grok-code-fast is already an amazing modal that covers 80% of usecases when coding, while being 10x faster and cheaper that similar models. I guess this is the same approach but for a more general public