Building the steroids for AI @ locul.ai x contextli.com (solving the accuracy & efficiency problem)

I feel like I’m frying my dopamine system with all this Claude/Codex work. Constantly monitoring and instructing, anticipating the work it’s going to produce. Then killing time waiting, consuming way more algo feed content than ever before. How do y’all tackle this?
272
31
1,559
84,206
Give work from your mobile via voice notes. Get reports of work done over email. Give feedback via mobile voice notes. Loop engineering basically. See link below if you need to set it up You ask it to mark all things that need you on desk to be "tagged" so you can just sit for half an hour on desk, take care of those - and then just work from your phoen mostly
1
2
Great post from Vibhor on what the new ChatGPT plugins are and why you should care. It was awesome to be a launch partner in this and give feedback along the way. Our revenue hasn’t stopped growing since we launched. I can’t stress enough how important this new surface is.
Hi, I’m Vibhor, Product Lead for Ecosystem at @OpenAI. I am going to use this space to share what we are doing to further the Ecosystem for our customers, and partners. DevDay feels like a good time to start. One of the best parts of DevDay was the deep collaboration with partners: @Adobe, @Atlassian, @Canva, @Figma, @Instacart, @MagicPathAI, @Shopify, Superpower, and @tldraw. While I went in excited I left with even more conviction in what we’re building together for our shared customers. It was also incredible to meet AI native partners and hear how our platform is empowering them build a sustainable business on ChatGPT 🎉 A conversation is often just the beginning of a task. You describe an idea, work through it, and eventually want something you can use: an edited image, a working prototype, a diagram you can keep refining. Getting there takes great tools and the people who know how to build them. Our ambition is to make those capabilities feel like a natural part of working with ChatGPT. At DevDay, we shared several steps in that direction. 1. Plugin Extensions help our partners bring interactive experiences to ChatGPT. Adobe is a good example. You can ask for an image edit, then adjust the details with hands-on controls. With tldraw, you and the agent can work on a shared canvas. MagicPath brings interactive prototypes into the conversation, with room to edit them yourself. These examples make the idea much more tangible: - Adobe: video.tv.adobe.com/v/3503940 - tldraw: x.lingyaoai.com/tldraw/status/21053958… - MagicPath: x.lingyaoai.com/skirano/status/2104994… - Figma: lnkd.in/gaqHXZjT - Shopify: lnkd.in/ghEXY5GT 2. A useful plugin can package skills, an MCP server, or both. That gives builders different ways to contribute: reusable workflow knowledge, access to live information, or actions that help someone finish a task. We introduced Plugin Creator and improvements to submissions, including automated checks and clearer feedback. Sites can also host supported plugins. There’s still work to do to make building and maintaining a plugin consistently straightforward. 3. We also announced improvements to plugin discovery, ranking and recommendations, in the directory and in conversations. This matters for both sides of the ecosystem. Users need to find something useful when they need it. Developers need a way for their work to reach those users. The goal is to make suggestions genuinely helpful, with good judgment about when they add value and when they get in the way. 4. MCP Events lets users ask ChatGPT to watch for supported events, such as a new message or a document comment, and respond when they arrive. It opens up another way for developers to make their tools useful beyond a single exchange. 5. ChatGPT Sites can now host MCP servers including plugin extensions! Create an MCP server with extensions, deploy the MCP server to Sites, turn the MCP server into a plugin, and install the plugin across platforms. There’s still plenty to do, especially around reliability, discovery and the developer experience. Thank you to the teams and partners doing this work with us. The future feels bright, and we’re just getting started. Onward!
7
16
349
41,639
"Revenue hasn't stopped growing" is doing a lot of heavy lifting with zero numbers attached. How are you even attributing the revenue, though? are users coming in through the plugin and converting inside your product, or is it mostly people seeing you in ChatGPT and then showing up later through search or direct?
5
Gemini 4.0 Pro is a BIG WIN for Google - It's 5x cheaper than Astra - even at 95% of Astra's performance, it's still frontier intelligence It has finally put Google ahead of open-source
99
46
1,384
69,465
Ahead of open-source on what though? Price per token, sure, but you can't self-host Gemini, can't fine-tune it on your own data, and Google can deprecate it whenever they feel like it. For a lot of builders that last 5% gap matters way less than owning the weights.
10
Now that AI can write any length, brevity is a luxury. People want short text. And I found a hack to make Claude or ChatGPT write this way. 1. The prompt: "Write in ASD-STE100. Be so brief every word you write costs you $1,000." 2. How to use it: At the beginning of your prompt, then just prompt it normally. A reminder that this is my favorite prompting format: "I want [goal] so that [success]. My situation is [context]. Ask me questions until you have enough context to start." 3. If you want it always: You need to go to your custom instructions. Think of them as instructions that apply to all your chats. → Claude: Go to Settings > Account > Instructions for Claude > Copy and paste the prompt I shared. → ChatGPT: Go to Settings > Personalization> Custom Instructions > Copy and paste the prompt. 4. You want to feel ahead, every week. I write 'How to AI', a newsletter to stay ahead of AI. It's twice a week, for free, on Sunday & Wednesday. Give me your email at how-to-ai.guide.
8
7
55
5,233
The $1,000 per word bit is going to bite people in the ass the second they ask for anything technical. Claude will happily drop the one caveat that mattered just to save a word. I'd rather cut the fluff myself than have it cut my edge cases :))
1
5
OpenAI is confirming the Astra 6.1 launch for next week A great time to also cut Astra’s price and push for adoption Models are going to self improve rapidly and become a lot more efficient
35
7
198
10,630
Cutting price the same week as a launch tells you exactly how scared they are of Anthropic right now.
7
Opus 5.5 Ultracode used $220 API Credits to generate this motion video in 8 hours. How do you think it looks?
1
3
91
built it for fun over the weekend -- got bored building SaaS last few weeks, need a change of pace
1
16
Opus 5.5 with UltraCode thinking has used up $200+ API tokens. Cooking a motion design video for 8 hours (still not done). I think I'm about to see THE BEST motion video of all time for playroomy.com/
2
1
50
Thoughts and prayers 🙏🏻 hope it comes out right after all that cooking
1
1
11
Haha, thanks -- it just got done with the mobile version. Have a look:
Opus 5.5 Ultracode used $220 API Credits to generate this motion video in 8 hours. How do you think it looks?
1
11
In 2008 Drew Houston had a file-syncing app called Dropbox and a waiting list stuck at about 5,000 people. He made one demo video, posted it where the right crowd hung out, and overnight the list jumped to 75,000. The product was the same one people had been ignoring the week before. Everyone remembers the video, but the part that mattered was who he made it for 🧵
1
1
2
40
my patience is running out
10
@DarioAmodei @sama heavy builders hit the top plan's limits every week. Give us a $500 to $1,000 plan: 10x the usage of the top tier Same models No weekly wall I'd pay today. So would @levelsio @marc_louvion @theo @theandreboso
24
GPT Sol and Astra as so obscenely good at browser use It is genuinely a new paradigm of how to do work I can basically guarantee you 2-3x productivity if you do this - install wispr. First thing first. Being able to tell your computer what to do is a game change instead of typing - install ChatGPT chrome plugin - brain dump everything you need done with wispr into codex with sol or Astra on medium - connect notion mcp - ask codex to creator a notion project tracker of everything you need to do from your brain dump - create 3 diff workspaces in codex. Name them agent 1, agent 2, agent 3 - go into each one and ask them to look at the notion board and figure out what parts of your the work it can do in the browser and to start - tell it to come to you for anything that requires decision making or is forward facing. Do not send any emails or anything that a client might see without your sign off And start Before you know it, you will see what the future looks like It will fundamentally change the way you see work
31
26
685
37,364
You can go one step further on this setup if you've got ~15 mins: Go complete hands-free where you basically dictate wispr-like on MOBILE, and your GPT/Claude session is checking in a 15min loop if a new note from you has come in. Also tell it to send you an email whenever anything gets done, so you can check on your phone -> then dictate in your phone and now you no longer gotta sit on your desk Cheers!
1
19
What’s one thing that’s missing in codex that you wish we had?
7,762
114
5,353
1,247,659
I'd love a Codex <-> ChatGPT memory. ChatGPT knows everything about me and my projects, but Codex doesn't.
29
558
35,272
You can use locul.ai for that; makes your memory layer far more robust and it ports over to claude code/desktop as well. Also uses your linkedin, dictation notes(wispr flow. Etc) and tons of other sources to build and auto update your second brain too
101
Replying to @zeq0r @thsottiaux
This can’t be stated enough. The amount of times I’m using ChatGPT/Work for research into what I’m trying to make it’d be so much easier for Codex to at least be able to share the context. It muddies the water too much to ask in the does chat directly for any research
1
5
1,100
You can use locul.ai for that; makes your memory layer far more robust and it ports over to claude code/desktop as well. Also uses your linkedin, dictation notes(wispr flow. Etc) and tons of other sources to build and auto update your second brain too
47
Since I went full-time on my startup there's a voice in my head that says "you should be working" whenever I watch a movie or go to bed. I took my first planned weekend off 16 months in. Does that voice ever go quiet for you?
34
In 2010 Kevin Systrom had an app called Burbn. Check-ins, plans, points, photos. It was running low on cash and people didn't stick around. So he deleted almost all of it. The stripped-down version had 25,000 people sign up on its first day. 18 months later Facebook paid about $1 billion for it, with 13 employees. My first startup did the exact opposite 🧵
3
2
252
Photo: @Scobleizer (Didn't know it was captured by him)
1
13