Twitter Trends April 23, 2026

Introducing GPT-5.5 A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its ...

Claude can now connect to more of the apps you use outside of work, including @Tripadvisor, @bookingcom, @resy, @Instacart, @Spotify, @audib...

We tried a new thing with NVIDIA to roll out Codex across a whole company and it was awesome to see it work. Let us know if you'd like to d...

Memory on Claude Managed Agents is now in public beta. Your agents can now learn from every session, using an intelligence-optimized memory...

1. We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a bi...

Images 2.0 really got over some important qualitative threshold for me that I didn't know existed.

Also, a ton of new Codex features coming soon! Fun little bundle w/the new model.

AI models are really good for teaching you how to build AI models.

I want to keep everyone updated on the details of the security investigation. The team performed an in-depth analysis to search for root ca...

Really excellent work by the inference team to serve this model so efficiently! To a significant degree, we have to become an AI inference ...

GStack is an open-source toolkit built by YC President & CEO @garrytan that turns Claude Code into an AI engineering team — with skills for ...

Cafe Cursor in Ghana

We have been testing GPT-5.5 in early access. Our evals show it’s the most capable model for people taking on complex builds with technical...

Dev bliss: Neovim in the center flanked by Kimi and Claude. Gotta love that TUI life 😍

This animation is one shotted on Codex with 5.4. Working on a skill that uses @QuiverAI to generate SVG characters then animates the charac...

New flow: 1. /triage command in Slack (#sandcastle) 2. Creates a discussion thread with agents for each issue in the repo needing triage 3....

I had early access to GPT-5.5. It is very good, especially the Pro version. Full writeup very shortly.

Yes... Claude Code is still better at design than GPT-5.5 But remember cmd-j opens the terminal. There you can run Claude Code directly ins...

You ever notice that some people say "lifestyle business" like it's some sorta slur?? One that pays well, gives you time back, and doesn't ...

So much of the business world is just surrounding valuable things that people want and need with the machinery that extracts higher margins ...

AI ate most of the software development lifecycle but maintaining live apps is still manual. “DevSecOps” is the new bottleneck as software...

It's crazy to me how much of software development is 'finding the right way to describe something that doesn't exist yet'. It's a language ...

Last week, we launched Gemini 3.1 TTS, our latest and best text-to-speech model. This new model introduces [awe] audio tags, an intuitive wa...

You can now use OpenAI OAuth to generate images with Hermes Agent! Update Hermes and use the ‘hermes tools’ command to configure the image ...

.@pmarca: "Don't keep a schedule" https://pmarchive.com/guide_to_personal_productivity.html

GPT-5.5-Pro on ARC-AGI (Verified) ARC-AGI-2: - Max: 82.2%, $10.76 - High: 84.6%, $10.51 GPT-5.5-Pro performs on par with GPT-5.5 on ARC-AG...

TrustMRR's newsletter just reached 10K readers 🎉 100% of the content is written by AI. I send a bunch of data to a cheap GPT model and ask...

How Anthropic’s product team moves faster than anyone else I sat down with @_catwu, Head of Product for Claude Code at @AnthropicAI, to get...

It looks like today's the day for GPT-5.5, we will probably get the livestream announcement shortly if so. Some of my friends had access and...

Codex having access to a great image model is such an underrated advantage over Claude.

One change over the last six months is that in every big company I talk, at least a few senior people absolutely get AI — they experiment (a...

Obsessed with turning entire research papers into conference posters extremely accurately with GPT Image 2! They really cooked with this mo...

Replit for Startups is live. Up to $25K in credits to build your product, ship fast, and scale. If you're building on Replit, this is for ...

Thanks @lennysan for a great conversation about how Claude Code maintains product velocity, how the product management role is shifting in t...

consumer products need utility for sure. but utility alone rarely creates pull. the products that actually earn repeated attention tend to f...

there are a bunch of people who talk to claude more than their mom, co-workers, cousins or best friends they just sit there prompt after p...

the iphone era app was a noun which loosely was a bounded surface you opened, performed an intent in, & closed. notes, reminders, clock, ca...

"Dude, the customer could not care less." Thiel Fellow @victorwboyd, who's building autonomous forklifts with the ultimate goal of being ab...

GPT 5.5 Experiment: Threejs: 25 trains, hills, switch views, change speeds, crashes, Crash Counter. Just a few prompts... idea from @peter...

not a single person i have ever spoken to uses gemini for coding. this is still very very weird. why is gemini so bad at coding when googl...

GStack /codex skill calls now fully explain what is going on. The non-verbal 200IQ Codex is now properly interpreted for us ADHD CEO types

Another NotebookLM papercut: ✅ Sharing notebooks with multiple people at once just got a whole lot easier. No more adding emails one by one...

Yeah 5.5 in Codex is really good at researching and creating spreadsheets, docs, and presentations... @OpenAI Please let me minimize the te...

A1Zap from @a1baseapi is the social network layer for AI apps. Starting with campuses, build apps and make interactive social content that s...

GPT-5.5 is here. Ben and I have had access for a bit. We have a lot of thoughts. 00:00:00 - Intro 00:01:34 - Vercel hack 00:05:38 - Kimi K2...

Among all the model release furor, important to note that people don’t need to switch providers or declare a winner every time a new model i...

With the Vantablack theme tweaked for true blacks in Omarchy 3.6, I've gotten the 16" Dell XPS OLED down to an incredible 2.2w idle. A pixel...

the per app model was a constraint of the pre llm era, when every domain required its own bespoke information architecture. ai collapses th...

Congrats to @mindfort on their $3M seed! They're building autonomous AI agents that find, validate, and patch vulnerabilities in productio...

You don’t have access to Mythos 🫵🤭 Doesn’t mean you can just sit around and wait. Replit published a whitepaper showing you can get sign...

Kinect (@trykinect) turns every e-commerce store into an AI-powered storefront that actually sells. As customers shop, online shopping assi...

Asendia AI builds AI recruiters for staffing agencies and enterprise companies. It clones their best recruiters into AI agents that match, s...

House lawmakers were given a demonstration by DHS yesterday where they were able to interact with jailbroken models. Open source will probab...

This is the big f**king deal. Cursor acquired for $60BN by xAI I sat down with @jasonlk and @rodriscoll to discuss the deal, along with th...

Unplugging completely! No WiFi and zero notifications. A great way to get deep focus on a project. Here is a walkthrough showing how to run...

We wanted to show you some stuff we've been hacking on: https://x.com/i/broadcasts/1pJdRbojDjVKW

GPT-5.5 solved my boring, hard problems (giant chunk of tech debt) and my fun problems (hacking into a bluetooth connected screen/speaker) w...

Dayjob is AI scheduling for waste trucks. They are at $496K in ARR with 12 customers. Congrats on the launch, @GHPos and Fred! https://www...

you should be spamming deep research right now... injecting context profiles of different ICPs, running market research for each segment, t...

Several of the tbench authors (myself included) will be presenting at ICLR tomorrow at 11:15am in Pavilion 4 P4-#5013. Come by to chat evals...

shipped 45,000 lines of code in the last 2 weeks. pre-Claude Code, this would've taken a senior dev 6 months. or a team of 2-3 ~2 months j...

Really impressed with gpt-image-2 for brand concepts, excited about how we can build on this.

Huscarl (http://huscarl.io) is an AI-native advisory firm for corporations buying insurance, helping them save 30% on annual premiums with z...

We interviewed 10 Thiel Fellows from the new 2026 class today. Here's what each one is working on:

in the early internet, design is what built trust that you weren't about to download an insane amount of viruses or have your credit card st...

Last night, I hosted a private dinner for founders and eng leaders of DevTools+AI companies in NYC 🗽 It was so cool to learn what Anthropi...

Turns out migration hardening matters a lot when you have 50k markdown files in your brain repo

Here’s my view on GPT-5.5, which I have been testing for a couple of weeks. It conducted not-bad social science research on its own, develo...

you need to develop an eye for when your context window is actually full... and i'm not talking about the percentage the model shows you m...

I'm going to say something shocking and borderline heretical (for me). I'm thinking about switching from Python to TypeScript for my next p...

opus 4.7 @ xhigh to plan and gpt 5.5 @ high to execute is the optimal setup, especially after today. i know opus to plan and gpt to execute...

Since Claude 3.5 Sonnet beat GPT-4 in June 2024, Claude has had the lead. GPT-5.5 just took it back. And Anthropic holding back Mythos is wh...

> spends the evening queuing up 15-20 tasks for AFK agents overnight > forgets to run the script that kicks them off > sigh

Looking back at the last 3 years. 2023 was effectively the last year of traditional developers. A few reflections: In 2023, we used Copil...

Thiel Fellow and cofounder of Opt32 @highpriestofswo on what's happening with CS majors: "I think there's going to need to be a reallocatio...

META plans to capture their employees keystrokes and mouse clicks while using sites like Google, LinkedIn, and Wikipedia to train their AI m...

A talk I gave a few weeks ago. Software fundamentals matter more than ever. Here's why: https://www.youtube.com/watch?v=v4F1gFy-hqg

fresh out of the lab... 🎥 webinars: the newest addition to the digital product suite 🎙️ podcast MCP: the first podcast host built for A...

My /github-triage skill in action

Researchers introduced Walrus, a transformer model that can predict what fluids (liquids, gases, and plasmas) will do across many physical d...

How many tokens per day do you use? Today I'm at 535M right now... is that a lot or average?

GPT 5.5 underperforms Opus 4.7 on SWE-Bench Pro. Couldn't find any reported SWE-Bench scores at all and an internal benchmark is reported in...

New: MCP support for Linear Agent Connect tools like Granola, Notion, PostHog, and many others to pull in context for creating issues, draf...

Thiel Fellow and cofounder of The Antifraud Company @alexkshieh says he's building the Palantir of fraud detection: "People have always bee...

This is a detailed walkthrough of how I run a 1-person VC-funded startup. Eng: @DevinAI Chief of Staff: @openclaw + @OpenAI Design: @Bret...

since the release of Sonnet 3.7, Anthropic was the ONLY company that seemed to understand how to leverage models horsepower... they mastere...

MTSLive Joins, xAI to Acquire Cursor, Mistral X xAI, GPT Images 2.0, Mythos Unauthorised Access https://x.com/i/broadcasts/1pKkOyereamKj

We’re building the economic infrastructure for AI. Tune in to hear everything we’re shipping for you and your agents from @patrickc @gaybric...

many models write a great plan and then get scared of it. @OpenAI's new GPT-5.5 just does the thing. here's what changed 👇

Rails World is coming to Austin, Texas in 2026. With last year selling out in mere minutes, we've made it bigger than ever with room for 1,2...

is there any way to make my firewall a bit more smart so when bots are spamming millions of requests from same place its auto denying them? ...

Bento grid prompt for GPT Image v2 Bento grids - handled well (the typography & brand precision are still random) Base prompt 👇

So I think social media CLIPPING might be one of the next big opportunities over the next two years. Not clipping directly, but building so...

missions - cleaning up droid - trying something new pi - organising + brainstorming living in cmux

Has anyone here (operating at scale) successfully automated their accounting/consolidated? DM me if yes.

Meet the New Thiel Fellows https://x.com/i/broadcasts/1wxWjaqypZDJQ

Chapter 1 of The Builder's Compass is live. "Old way: documentation to prevent failure. New way: prototyping to demonstrate success." AI has...

The average online course completion rate is 13%. Over half never open the material. You can create a great course but what does it matter...

Introducing Tab Snitch Liven up your screenshots and make co-workers uncomfortable with errant open tabs. http://tab.wesbos.com

I used to think in-app surveys didn’t work. Then we analyzed 4.2M responses across nearly 6,000 surveys. The problem was never surveys. Th...