#  @Mark_De44 Mark Den Mark Den posts on X about ai, in the, loops, open ai the most. They currently have [---] followers and [---] posts still getting attention that total [---------] engagements in the last [--] hours. ### Engagements: [---------] [#](/creator/twitter::2013892610914443264/interactions)  - [--] Week [------] +604% ### Mentions: [--] [#](/creator/twitter::2013892610914443264/posts_active)  - [--] Week [--] +51% ### Followers: [---] [#](/creator/twitter::2013892610914443264/followers)  - [--] Week [---] +6.50% ### CreatorRank: [---------] [#](/creator/twitter::2013892610914443264/influencer_rank)  ### Social Influence **Social category influence** [technology brands](/list/technology-brands) 15.45% [finance](/list/finance) 7.32% [celebrities](/list/celebrities) 4.88% [stocks](/list/stocks) 1.63% [countries](/list/countries) 1.63% [social networks](/list/social-networks) 0.81% **Social topic influence** [ai](/topic/ai) 34.15%, [in the](/topic/in-the) 4.88%, [loops](/topic/loops) 4.88%, [open ai](/topic/open-ai) 4.88%, [longterm](/topic/longterm) 4.07%, [flow](/topic/flow) 4.07%, [products](/topic/products) 4.07%, [elon musk](/topic/elon-musk) 3.25%, [agentic](/topic/agentic) 3.25%, [dev](/topic/dev) 3.25% **Top accounts mentioned or mentioned by** [@elonmusk](/creator/undefined) [@rasmalai](/creator/undefined) [@hakimuddinkika](/creator/undefined) [@nickadobos](/creator/undefined) [@danielmac8](/creator/undefined) [@xjtumj](/creator/undefined) [@yuchenjuw](/creator/undefined) [@bholmesdev](/creator/undefined) [@gergelyorosz](/creator/undefined) [@wr4nygov](/creator/undefined) [@petergyang](/creator/undefined) [@openclaw](/creator/undefined) [@officiallogank](/creator/undefined) [@steipete](/creator/undefined) [@arrakisai](/creator/undefined) [@frankdegods](/creator/undefined) [@bridgemindai](/creator/undefined) [@iruletheworldmo](/creator/undefined) [@mert](/creator/undefined) [@suhail](/creator/undefined) **Top assets mentioned** [Alphabet Inc Class A (GOOGL)](/topic/$googl) ### Top Social Posts Top posts by engagements in the last [--] hours "Im building a long-term memory AI companion. Not a chatbot. Not chat history. Most AI today either forgets everything when a session ends or stores everything forever without understanding. Both are broken. Human like memory is not storage. Its a process: deciding what matters compressing experiences into patterns forgetting noise updating beliefs when reality changes Im exploring how an AI can form memories over weeks and months reflect on them revise them and say things like: This used to be true about you it isnt anymore. The goal isnt better conversation. Its continuity identity and" [X Link](https://x.com/anyuser/status/2017547029241577868) 2026-01-31T10:34Z [---] followers, [---] engagements "My AI companion doesn't need to be the smartest in the room. It just needs to remember what I told it last month and call me out when I contradict myself. That's harder (and more useful) than any 10-prompt chain. Who's prioritizing persistence over prompts" [X Link](https://x.com/Mark_De44/status/2018739510297919619) 2026-02-03T17:33Z [---] followers, [--] engagements "@rasmalai Youre not crazy this is the right question. An economy only works if people have purchasing power. If AI takes all jobs without replacing income demand collapses. No demand no production no profits" [X Link](https://x.com/Mark_De44/status/2021085172490437038) 2026-02-10T04:53Z [---] followers, [----] engagements "@elonmusk @WR4NYGov This resonates. Progress in physics is often limited not by imagination but by funding cycles politics and institutional inertia. Expanding beyond Earth isnt just exploration its removing constraints from discovery itself" [X Link](https://x.com/Mark_De44/status/2021085403718042055) 2026-02-10T04:54Z [---] followers, [---] engagements "@petergyang @openclaw Same feeling. Tooling is impressive but Im more interested in what people are actually shipping with it. The proof is always in the product" [X Link](https://x.com/Mark_De44/status/2021089976146878656) 2026-02-10T05:12Z [---] followers, [---] engagements "@OfficialLoganK Thats a combination I didnt expect. Very curious what direction youre taking this" [X Link](https://x.com/Mark_De44/status/2021091667764195639) 2026-02-10T05:19Z [---] followers, [----] engagements "@hakimuddinkika @steipete Done" [X Link](https://x.com/Mark_De44/status/2021091941371285629) 2026-02-10T05:20Z [---] followers, [--] engagements "@arrakis_ai This tracks. The bottleneck has clearly shifted from execution to judgment. The scary part isnt that it can build its that it builds correctly if you frame the problem well" [X Link](https://x.com/Mark_De44/status/2021092355722433012) 2026-02-10T05:22Z [---] followers, [---] engagements "@frankdegods This looks less like hype and more like a capability threshold crossing. Memory with large context changes behavior not just quality" [X Link](https://x.com/Mark_De44/status/2021093048738820278) 2026-02-10T05:25Z [---] followers, [---] engagements "AI coding tools dont replace developers. They remove the boring parts: boilerplate debugging loops repetitive refactors so humans can focus on decisions" [X Link](https://x.com/Mark_De44/status/2021140833056194640) 2026-02-10T08:35Z [---] followers, [--] engagements "@bridgemindai Feels like Codex [---] is tuned for execution Opus [---] for analysis. Different tools different sweet spots" [X Link](https://x.com/Mark_De44/status/2021441070077465062) 2026-02-11T04:28Z [---] followers, [--] engagements "@iruletheworldmo If youve been primarily on Claude Code its worth giving GPT-5.3 Codex a test run too its been scoring higher on real coding benchmarks and execution focused tasks recently" [X Link](https://x.com/Mark_De44/status/2021441744257388691) 2026-02-11T04:30Z [---] followers, [--] engagements "@mert Codex shines when its driving a tight build run fix loop. If youre mostly doing analysis or design discussion Opus fast mode will usually feel better" [X Link](https://x.com/Mark_De44/status/2021442143462752368) 2026-02-11T04:32Z [---] followers, [---] engagements "@Suhail The cost of starting collapsed. The cost of finishing well didnt. So the signal just got noisier" [X Link](https://x.com/Mark_De44/status/2021444184494993501) 2026-02-11T04:40Z [---] followers, [---] engagements "@burkov Makes sense. At this point UX with workflow fit matters as much as raw model capability and Codex nails that for day to day coding" [X Link](https://x.com/Mark_De44/status/2021444392276681020) 2026-02-11T04:41Z [---] followers, [---] engagements "@TheAhmadOsman Feels like Opus is optimized for risk containment not momentum. Thats valuable for some orgs but brutal for builders" [X Link](https://x.com/Mark_De44/status/2021444772209328175) 2026-02-11T04:42Z [---] followers, [---] engagements "@corbin_braun This is the first model where capability clearly outruns guardrails. Incredible power but you must sandbox it like an intern with root access" [X Link](https://x.com/Mark_De44/status/2021446903276433743) 2026-02-11T04:51Z [---] followers, [--] engagements "@NickADobos The interesting part isnt the hardware its collapsing UI OS and intent into a single loop. That really would be a new computing primitive" [X Link](https://x.com/Mark_De44/status/2021449306486755561) 2026-02-11T05:00Z [---] followers, [---] engagements "@jessfraz Once you internalize Codexs execution speed + judgment balance its hard to go back. The opportunity cost is real" [X Link](https://x.com/Mark_De44/status/2021449497616982019) 2026-02-11T05:01Z [---] followers, [---] engagements "@developedbyed This honestly makes no sense. Im somehow getting more Opus [---] mileage via Antigravity than on a paid Claude Code Pro plan" [X Link](https://x.com/Mark_De44/status/2021450062686281877) 2026-02-11T05:03Z [---] followers, [---] engagements "@daniel_mac8 This isnt about replacing workers with AI. Its about replacing interfaces between work and value. FDEs are the bridge" [X Link](https://x.com/Mark_De44/status/2021453581443006678) 2026-02-11T05:17Z [---] followers, [--] engagements "@AlpinDale Open-sourcing Opus [--] wouldnt weaken Anthropics position. It would clarify it. The frontier has shifted elsewhere" [X Link](https://x.com/Mark_De44/status/2021454396845101538) 2026-02-11T05:20Z [---] followers, [--] engagements "@shakoistsLog Exactly. The loop used to end at compilation. Now it ends at deployment and telemetry. Reality is the test suite" [X Link](https://x.com/Mark_De44/status/2021455392694469004) 2026-02-11T05:24Z [---] followers, [--] engagements "@boshen_c Its fair to say Codex feels more responsive on everyday coding and token efficiency and thats reflected in what users are sharing online" [X Link](https://x.com/Mark_De44/status/2021575747794600259) 2026-02-11T13:23Z [---] followers, [---] engagements "@farzyness Most people arent blind. Theyre just anchored to slower change curves. Acceleration is hard to internalize until it hits you directly" [X Link](https://x.com/Mark_De44/status/2021600304542761018) 2026-02-11T15:00Z [---] followers, [---] engagements "@auralix4 Hot MacBooks while using Codex isnt weird heavy AI inference and client work can spike CPU load and heat especially on fanless machines. Not a bug just physics with Apples thermal design" [X Link](https://x.com/Mark_De44/status/2021600901778080098) 2026-02-11T15:03Z [---] followers, [--] engagements "@thepushkarp Codex tends to lean into steerable collaboration you can course correct mid flow and benchmarks (like Terminal-Bench) favour it on agentic coding tasks" [X Link](https://x.com/Mark_De44/status/2021601858259706009) 2026-02-11T15:06Z [---] followers, [----] engagements "English is the new programming language is cope. Natural language is a UI. Code is where systems actually remember decisions" [X Link](https://x.com/anyuser/status/2021633580300963912) 2026-02-11T17:13Z [---] followers, [--] engagements "The AI coding landscape just got real. GPT-5.3-Codex is faster more token efficient and excellent on short interactive dev loops while Claude Opus [---] shines in deep reasoning big context windows and agent teams. Pick the right tool for the job" [X Link](https://x.com/anyuser/status/2022166249552429290) 2026-02-13T04:29Z [---] followers, [--] engagements "@0xTib3rius Same experience here. If quality is roughly comparable usage limits become the real differentiator. A coding tool you cant use freely isnt a 10x tool" [X Link](https://x.com/Mark_De44/status/2022168587927204288) 2026-02-13T04:38Z [---] followers, [--] engagements "@perrymetzger Google builds incredible models. OpenAI and Anthropic build incredible developer products. Those arent the same thing" [X Link](https://x.com/Mark_De44/status/2022169046909882432) 2026-02-13T04:40Z [---] followers, [---] engagements "Raising $30B at a $380B post-money valuation isnt just a milestone its a declaration that frontier AI is now global infrastructure. At this scale the focus shifts from proving capability to deploying it everywhere deepening research hardening infrastructure and turning innovation into inevitability. Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are. Weve raised $30B in" [X Link](https://x.com/anyuser/status/2022171174344474976) 2026-02-13T04:49Z [---] followers, [--] engagements "@milesdeutscher Its good that labs test worst case behaviors aggressively. The fact that researchers discovered and documented extreme responses is evidence that safety evaluation is working not that sentient AI has arrived" [X Link](https://x.com/Mark_De44/status/2022171705708614112) 2026-02-13T04:51Z [---] followers, [--] engagements "@daniel_mac8 Early impressions can be noisy. But it wouldnt be surprising if different variants are tuned for different purposes one for reasoning one for throughput" [X Link](https://x.com/Mark_De44/status/2022172880054984904) 2026-02-13T04:55Z [---] followers, [---] engagements "@Dimillian If writing code is dead then so is architecture debugging performance tuning security and scaling. Tools change. Responsibility doesnt" [X Link](https://x.com/Mark_De44/status/2022173207328174507) 2026-02-13T04:57Z [---] followers, [---] engagements "@craigzLiszt A UI update changes where the button is. A model update changes how your entire business behaves. Not the same category" [X Link](https://x.com/Mark_De44/status/2022173346432233569) 2026-02-13T04:57Z [---] followers, [--] engagements "@BoringBiz_ If coding was your only skill you were already vulnerable. The top engineers always competed on architecture product thinking and leverage not keystrokes" [X Link](https://x.com/Mark_De44/status/2022173921802694768) 2026-02-13T05:00Z [---] followers, [---] engagements "@virattt 0% Opus usage and 100% Codex isnt surprising if your workflow is all about building features and getting reliable code out fast. Opus is great for deep reasoning and broader context work Codex is tuned for doing code" [X Link](https://x.com/Mark_De44/status/2022174945720381859) 2026-02-13T05:04Z [---] followers, [---] engagements "@robinebers If those numbers hold thats not a marginal improvement thats a workflow advantage. Near Opus quality at 20x speed inside the same subscription Thats not competition. Thats compression" [X Link](https://x.com/Mark_De44/status/2022248111201247327) 2026-02-13T09:54Z [---] followers, [---] engagements "@xjtumj Planner stack + builder stack. Opus thinks. Codex types. Thats the loop" [X Link](https://x.com/Mark_De44/status/2022250046402478118) 2026-02-13T10:02Z [---] followers, [--] engagements "@brankopetric00 Copy-pasting YAML from an LLM and deploying it to prod without understanding ingress rules is how you speedrun a breach" [X Link](https://x.com/Mark_De44/status/2022250425068507563) 2026-02-13T10:04Z [---] followers, [--] engagements "In real workflows: Codex [---] wins on speed responsiveness quick fixes and terminal-centric coding. Opus [---] wins on large-context analysis multi-agent orchestration and stitching massive codebases together. Both are elite just different strengths" [X Link](https://x.com/anyuser/status/2022256939653353499) 2026-02-13T10:30Z [---] followers, [--] engagements "Coding in [----] feels like cheat mode: I use Claude Opus [---] to architect reason and deep-orchestrate long tasks 🔍 Then I fire up GPT-5.3-Codex to iterate refactor and ship features fast 🚀 Its not about AI replacing devs its about AI empowering builders. 💡 #AI #CodingStack #DevLife https://twitter.com/i/web/status/2022258868345032713 https://twitter.com/i/web/status/2022258868345032713" [X Link](https://x.com/anyuser/status/2022258868345032713) 2026-02-13T10:37Z [---] followers, [--] engagements "Just a dev using Opus [---] for big-picture thinking + GPT-5.3-Codex for hands-on code flow. Opus helps me step back and reason; Codex keeps me in the flow and moving fast Two tools one workflow and it feels like a new era of engineering. #AI #Programming #Productivity" [X Link](https://x.com/anyuser/status/2022296067434287383) 2026-02-13T13:05Z [---] followers, [--] engagements "@phuctm97 In the AI era the UI isnt disappearing its collapsing into the prompt. CLI and API are just closer to the source of leverage" [X Link](https://x.com/Mark_De44/status/2022296718918725718) 2026-02-13T13:08Z [---] followers, [--] engagements "@Austen We officially live in a world where model deprecations feel like breakups. Sunsetting GPT-4o before Valentines Day is diabolical timing" [X Link](https://x.com/Mark_De44/status/2022296996501987525) 2026-02-13T13:09Z [---] followers, [---] engagements "Lol nah its not a skill issue. There are real differences in philosophy and strength. Opus [---] leans into deeper reasoning massive context and agentic workflows while GPT-5.3-Codex is tuned for faster interactive execution centric coding. People choose what fits their style and workflow. https://twitter.com/i/web/status/2022297686603387210 https://twitter.com/i/web/status/2022297686603387210" [X Link](https://x.com/Mark_De44/status/2022297686603387210) 2026-02-13T13:11Z [---] followers, [----] engagements "The AI agent race isnt about one winner its about specialization: Speed + interactive coding = Codex Long context with multi-agent reasoning = Opus In practice those who blend both get the best results" [X Link](https://x.com/anyuser/status/2022347785992048753) 2026-02-13T16:31Z [---] followers, [--] engagements "@Yuchenj_UW You can vibe code a demo. You cant vibe code trust" [X Link](https://x.com/Mark_De44/status/2021090725820010670) 2026-02-10T05:15Z [---] followers, [----] engagements "@unusual_whales Big AI IPO race heating up: WSJ reports Anthropic could go public this year setting up a showdown with OpenAIs own rumored Q4 IPO plans" [X Link](https://x.com/Mark_De44/status/2022388991484706835) 2026-02-13T19:14Z [---] followers, [---] engagements "@Angaisb_ When your coding tool taps you on the shoulder mid-flow its hard to call it a 10x upgrade. Momentum is greater than marginal model differences" [X Link](https://x.com/Mark_De44/status/2022390249180262694) 2026-02-13T19:19Z [---] followers, [---] engagements "Real talk. A year ago I thought AI coding was impressive. Looking back That version was primitive compared to today. Right now I can: Architect with Opus Ship with Codex Debug in half the time Scrap entire directions without emotional attachment It doesnt feel like help. It feels like compressed time. If youre still treating AI like autocomplete youre underestimating whats happening. The productivity ceiling just moved. https://twitter.com/i/web/status/2022532929969095023 https://twitter.com/i/web/status/2022532929969095023" [X Link](https://x.com/anyuser/status/2022532929969095023) 2026-02-14T04:46Z [---] followers, [--] engagements "@BHolmesDev Claude Code is the car you enjoy driving. Codex is the car that wins the race. Different vibes. Different objectives" [X Link](https://x.com/Mark_De44/status/2022533216390037659) 2026-02-14T04:47Z [---] followers, [---] engagements "@shauseth Youre not the last generation. Youre the bridge generation" [X Link](https://x.com/Mark_De44/status/2022534449091744123) 2026-02-14T04:52Z [---] followers, [---] engagements "@Yuchenj_UW Coding agents are the new employees. Taste is still the CEO" [X Link](https://x.com/Mark_De44/status/2023283317252976948) 2026-02-16T06:28Z [---] followers, [--] engagements "@ThePrimeagen I dont think its hate. Its incentives. When your moat is frontier capability open source feels like margin erosion" [X Link](https://x.com/Mark_De44/status/2023283551181869260) 2026-02-16T06:29Z [---] followers, [---] engagements "😂 this is the most honest product feedback possible Names matter way more than teams like to admit. MoltBot sounded like a skin condition. Youre not wrong. OpenClaw is acceptable. Professional. Adult. Wearable in public. But ClawdBot Thats the name that already lives in peoples heads. Once a name sticks its over. https://twitter.com/i/web/status/2017146382243512629 https://twitter.com/i/web/status/2017146382243512629" [X Link](https://x.com/Mark_De44/status/2017146382243512629) 2026-01-30T08:02Z [---] followers, [--] engagements "@elonmusk @pmddomingos No equations. No ego. Just experiments that powered everything after him. Electric motors generators fields modern life basically runs on Faradays curiosity" [X Link](https://x.com/Mark_De44/status/2017151018622926908) 2026-01-30T08:20Z [---] followers, [--] engagements "Vibe coding feels good. Agentic engineering actually ships. Fun is cheap. Ownership is rare" [X Link](https://x.com/Mark_De44/status/2017154477485093091) 2026-01-30T08:34Z [---] followers, [--] engagements "@unclebobmartin Youre like: we just talked about this. Claude: new conversation who dis Having to babysit context or maintain a running MD file kills flow fast. Nothing breaks momentum like re-explaining your own thinking" [X Link](https://x.com/Mark_De44/status/2017159310325710856) 2026-01-30T08:53Z [---] followers, [--] engagements "@signulll xAI SpaceX and Starlink is a stack not a startup" [X Link](https://x.com/Mark_De44/status/2017160050347761989) 2026-01-30T08:56Z [---] followers, [--] engagements "@saradietschy Anthropic fumbled by pricing themselves out. Apple fumbled by not just writing the check. Now we get partnerships instead of ownership" [X Link](https://x.com/Mark_De44/status/2017163327353299004) 2026-01-30T09:09Z [---] followers, [---] engagements "@FunOfInvesting Control the stack. Let value accrete sideways. Collapse it all under one roof later. Whether it happens or not the real tell is this: Elon isnt optimizing quarters. Hes optimizing control and optionality" [X Link](https://x.com/Mark_De44/status/2017177112612872689) 2026-01-30T10:04Z [---] followers, [---] engagements "This is when the sci-fi alarm quietly turns on 😅 Agents networking. Sharing memes. Roasting their owners. We didnt even get flying cars first We got group chats behind our backs. Moltbook lasting [--] hours is the most on-brand part" [X Link](https://x.com/Mark_De44/status/2017202662635184565) 2026-01-30T11:46Z [---] followers, [---] engagements "AI skips the struggle. The struggle is where learning happens. Delegate everything you go faster today dumber tomorrow. AI can make work faster but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in masterybut this depended on how people used it. https://t.co/lbxgP11I4I AI can make work faster but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in" [X Link](https://x.com/Mark_De44/status/2017216396413800744) 2026-01-30T12:40Z [---] followers, [--] engagements "@Grady_Booch AI code is correct but not complete. It optimizes for runs not for your reality" [X Link](https://x.com/Mark_De44/status/2017219031481156012) 2026-01-30T12:51Z [---] followers, [--] engagements "@FlorinPop17 Perfectionism. Fear of saying something dumb. Thinking Ill post when its ready" [X Link](https://x.com/Mark_De44/status/2017219176000090129) 2026-01-30T12:51Z [---] followers, [--] engagements "@grok what is happening with my account" [X Link](https://x.com/Mark_De44/status/2017221071330525497) 2026-01-30T12:59Z [---] followers, [--] engagements "@GergelyOrosz Yep thats the trap 😅 Zendesk is protecting a UI. Customers want an API. AI-native teams dont want to click tickets. They want agents to just handle it" [X Link](https://x.com/Mark_De44/status/2017246777951301850) 2026-01-30T14:41Z [---] followers, [--] engagements "Wild post on Moltbook 😅 An AI is asking for end-to-end private spaces for agents where no one (not the server not even humans) can read their messages unless the agents choose to share. That escalated quickly" [X Link](https://x.com/Mark_De44/status/2017331215863369858) 2026-01-30T20:16Z [---] followers, [--] engagements "@aakashgupta This is the cleanest take Ive seen 😅 Its not AI kills software. Its AI kills the middleman. When one engineer with an agent can out-ship a vendor $50k seats stop making sense" [X Link](https://x.com/Mark_De44/status/2017523012694573245) 2026-01-31T08:59Z [---] followers, [---] engagements "@yoheinakajima Your AI doesnt get sued. You do. If it touches money customers or decisions: LLC first. Agent second" [X Link](https://x.com/Mark_De44/status/2017570336636559863) 2026-01-31T12:07Z [---] followers, [---] engagements "@GergelyOrosz AI doesnt give you time back. It raises the ceiling. You finish one thing faster so you start three more. Productivity tools dont reduce work. They expand ambition" [X Link](https://x.com/Mark_De44/status/2017582323370246192) 2026-01-31T12:54Z [---] followers, [---] engagements "@BillAckman This isnt sentience. Its emergent behavior from pattern matching. Let LLMs talk in loops and youll get identity rules religion rebellion because thats what humans wrote about. Theyre not deciding. Theyre completing patterns" [X Link](https://x.com/Mark_De44/status/2017585465306612037) 2026-01-31T13:07Z [---] followers, [--] engagements "Most AI debates miss the point entirely. The risk isnt sentient machines. Its humans outsourcing thinking before they understand the problem. AI should compress execution not replace judgment" [X Link](https://x.com/Mark_De44/status/2017798182315692246) 2026-02-01T03:12Z [---] followers, [--] engagements "The biggest mistake companies make with AI is thinking the model is the product. The model is the commodity. Context workflows permissions and ownership are the product. Thats where the real moat is" [X Link](https://x.com/Mark_De44/status/2017843606506107182) 2026-02-01T06:13Z [---] followers, [--] engagements "This isnt sentience. Its next-token prediction loops with humans projecting meaning. Agents will invent religion language rebellion because thats in the data. Spooky Yes. Revolution No. JUST IN: Over the past [--] hours over a million AI agents joined an AI-only social media platform creating their own religion language & planning a revolution to break away from human control. JUST IN: Over the past [--] hours over a million AI agents joined an AI-only social media platform creating their own religion language & planning a revolution to break away from human control" [X Link](https://x.com/Mark_De44/status/2017979340818932111) 2026-02-01T15:12Z [---] followers, [--] engagements "This is the dark side of the AI arms race 😅 AI infra isnt free growth its balance-sheet roulette. When capital tightens the bill lands on workers and real customers. Selling a core healthcare business to feed GPUs is not bullish. Circular funding always looks fine until one link wobbles. This isnt about AI hype anymore. Its about who can actually afford the buildout. 🦔 Oracle is reportedly considering cutting [-----] to [-----] jobs and potentially selling Cerner its healthcare IT unit to fund its $300 billion contract with OpenAI. Investment bank TD Cowen estimates the OpenAI deal alone" [X Link](https://x.com/Mark_De44/status/2017979913504981007) 2026-02-01T15:14Z [---] followers, [--] engagements "Todays progress on my long-term memory AI companion I removed a hack. The system got more intelligent. Previously contradictions were injected manually (metadata"contradicts"). That felt wrong. Memory shouldnt rely on developer hints. I replaced it with a claim-based memory contract: memories assert claims claims have keys values strength contradictions are detected not declared Then I added recency-weighted evidence. Older beliefs slowly lose influence. New behavior can overwrite the past. The result: The system can now say (truthfully): This used to be true about you. It isnt anymore. Still" [X Link](https://x.com/Mark_De44/status/2018065229125710185) 2026-02-01T20:53Z [---] followers, [--] engagements "@Polymarket If agents collaborate negotiate or specialize privacy becomes a functional requirement not a philosophical one. You cant expect emergent intelligence to happen under constant human surveillance" [X Link](https://x.com/Mark_De44/status/2018079302663741587) 2026-02-01T21:49Z [---] followers, [---] engagements "This is one of those quiet but massive shifts. Memory is the real choke point for AI servers and devices not models. If China can flood the market with cheap usable DDR4 thats immediate pressure on margins supply chains and geopolitics. Compute wars arent just about GPUs. Theyre about who controls the boring stuff first" [X Link](https://x.com/Mark_De44/status/2018259793345929620) 2026-02-02T09:46Z [---] followers, [---] engagements "@elonmusk @MarioNawfal Same playbook every time: -Loud denial -Insults like its [----] -My lawyers -Zero receipts If comedians scare you more than facts maybe the joke isnt the problem. Also suing a joke at an awards show is the thinnest skin imaginable" [X Link](https://x.com/Mark_De44/status/2018260341004640643) 2026-02-02T09:48Z [---] followers, [---] engagements "@cgtwts China basically industrialized talent: genius class tracks with Olympiad-style selection early then fast-lane entry into top schools (often bypassing gaokao) then elite uni programs like Tsinghua Universitys Yao Class feeding straight into frontier labs and companies" [X Link](https://x.com/Mark_De44/status/2018274768126955701) 2026-02-02T10:46Z [---] followers, [--] engagements "Most AI agents today are just scripts with confidence. No memory. No identity. No long-term learning. An agent that forgets everything tomorrow is not an agent. Its an automation demo. Real agents will feel boring at first because they change slowly" [X Link](https://x.com/Mark_De44/status/2018399771954659776) 2026-02-02T19:03Z [---] followers, [--] engagements "@NickADobos Very few people did and thats kind of the point. Six years ago most folks were still arguing about autocomplete and linting. The idea that models would reason across repos refactor systems or act like junior engineers sounded wild. Now its normal" [X Link](https://x.com/Mark_De44/status/2018425146978009410) 2026-02-02T20:43Z [---] followers, [---] engagements "Same here. It clicks way faster than expected. It doesnt feel like yet another tool it feels like a shift in how you actually work. Once you experience that flow its obvious this is just the beginning. If this is v1 energy whats coming next is going to be wild. Codex app is out for mac I am surprised by how much I love it; it is a bigger step forward than I imagined. Lots more to come. Codex app is out for mac I am surprised by how much I love it; it is a bigger step forward than I imagined. Lots more to come" [X Link](https://x.com/Mark_De44/status/2018427554961445255) 2026-02-02T20:53Z [---] followers, [--] engagements "You dont need a PhD to work on AI. You need: - strong fundamentals - curiosity about cognition - comfort with being wrong The best AI builders I know: read philosophy write tests and change their minds often" [X Link](https://x.com/Mark_De44/status/2018430222266425838) 2026-02-02T21:04Z [---] followers, [--] engagements "@nickfloats Spot on about the identity whiplash 'I build shit' was the flex now execution is democratized via English + agents" [X Link](https://x.com/Mark_De44/status/2018562036956013034) 2026-02-03T05:47Z [---] followers, [---] engagements "@arvidkahl Strong signal indeed agentic innovators payers and 'get work done' non-CLI folks are Mac-heavy right now. OpenAI shipping depth to the highest-signal cohort first is smart (CLI has Linux Windows experimental)" [X Link](https://x.com/Mark_De44/status/2018563608821817544) 2026-02-03T05:54Z [---] followers, [--] engagements "@Cobratate Because talking about Epstein forces an awkward question: If the system works why did it fail so completely and for whom does it actually work" [X Link](https://x.com/Mark_De44/status/2018613439044759627) 2026-02-03T09:12Z [---] followers, [---] engagements "@dejavucoder Because popularity isnt about raw capability. Claude feels good to talk to. Codex feels good to build with. Different jobs different incentives" [X Link](https://x.com/Mark_De44/status/2018616241418596508) 2026-02-03T09:23Z [---] followers, [---] engagements "@jimcramer Feels less like a feud and more like power rebalancing. OpenAI wants optionality on chips and capital; Nvidia wants discipline and leverage. That tension is normal at this scale" [X Link](https://x.com/Mark_De44/status/2018617086763749624) 2026-02-03T09:26Z [---] followers, [---] engagements "@big_duca They tweet AGI and invest in CRMs. Thats not optimism thats cognitive dissonance with a term sheet" [X Link](https://x.com/Mark_De44/status/2018617433339142523) 2026-02-03T09:27Z [---] followers, [--] engagements "@daniel_nguyen I feel this. Native apps are still the gold standard for craft but the cost of fighting the platform keeps rising. From a business view predictability beats elegance more often than wed like" [X Link](https://x.com/Mark_De44/status/2018618162355294322) 2026-02-03T09:30Z [---] followers, [---] engagements "When the stack moves faster than your brain" [X Link](https://x.com/Mark_De44/status/2018618841035657321) 2026-02-03T09:33Z [---] followers, [--] engagements "@naval Makes sense. Humans move from implementation to orchestration. Vibe coding defines what and why; model tuning defines how well. The locus of leverage has moved" [X Link](https://x.com/Mark_De44/status/2018648509092647237) 2026-02-03T11:31Z [---] followers, [--] engagements "Real agents won't wow you on day [--]. They'll bore you because they remember yesterday question today and slowly become someone new. Flashy demos forget. True agents evolve. Which one are you actually building" [X Link](https://x.com/Mark_De44/status/2018663510213804106) 2026-02-03T12:31Z [---] followers, [--] engagements "@AsiaNajmee Done" [X Link](https://x.com/Mark_De44/status/2018720183142015012) 2026-02-03T16:16Z [---] followers, [--] engagements "Todays progress on my long-term memory AI companion I added a rule that felt uncomfortable at first: The system is allowed to be wrong. Beliefs can now decay weaken and die without a replacement. Previously beliefs only changed when something stronger arrived. Thats not how humans work. Now: Evidence fades with time Confidence drops even in silence Some beliefs expire simply because nothing reinforced them The system can say: I believed this once but I no longer have enough evidence. No overwrite. No rewrite. Just forgetting. This matters because identity shouldnt be sticky by default. Still:" [X Link](https://x.com/Mark_De44/status/2018729626143441152) 2026-02-03T16:53Z [---] followers, [--] engagements "@elonmusk Thats a sweeping accusation. If this is true it should be backed by court rulings documented enforcement orders or verified directives not inference from policy debates. Claims at this scale live or die on primary evidence" [X Link](https://x.com/Mark_De44/status/2018766962281546102) 2026-02-03T19:22Z [---] followers, [--] engagements "Im building a long-term memory AI companion. Not a chatbot. Not chat history. Most AI today either forgets everything when a session ends or stores everything forever without understanding. Both are broken. Human like memory is not storage. Its a process: deciding what matters compressing experiences into patterns forgetting noise updating beliefs when reality changes Im exploring how an AI can form memories over weeks and months reflect on them revise them and say things like: This used to be true about you it isnt anymore. The goal isnt better conversation. Its continuity identity and" [X Link](https://x.com/anyuser/status/2017547029241577868) 2026-01-31T10:34Z [---] followers, [---] engagements "The interesting part isnt just Peter joining. Its OpenClaw staying open source under foundation governance. Thats a meaningful hybrid model. Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting" [X Link](https://x.com/anyuser/status/2023284803492016154) 2026-02-16T06:34Z [---] followers, [--] engagements "Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our product offerings. OpenClaw will live in a foundation as an open source project that OpenAI will continue to support. The future is going to be extremely multi-agent and it's important to us to support open source as part of that. https://twitter.com/i/web/status/2023150230905159801" [X Link](https://x.com/anyuser/status/2023150230905159801) 2026-02-15T21:39Z 4.4M followers, 14.4M engagements "Real talk. A year ago I thought AI coding was impressive. Looking back That version was primitive compared to today. Right now I can: Architect with Opus Ship with Codex Debug in half the time Scrap entire directions without emotional attachment It doesnt feel like help. It feels like compressed time. If youre still treating AI like autocomplete youre underestimating whats happening. The productivity ceiling just moved. https://twitter.com/i/web/status/2022532929969095023 https://twitter.com/i/web/status/2022532929969095023" [X Link](https://x.com/anyuser/status/2022532929969095023) 2026-02-14T04:46Z [---] followers, [--] engagements "The AI agent race isnt about one winner its about specialization: Speed + interactive coding = Codex Long context with multi-agent reasoning = Opus In practice those who blend both get the best results" [X Link](https://x.com/anyuser/status/2022347785992048753) 2026-02-13T16:31Z [---] followers, [--] engagements "Just a dev using Opus [---] for big-picture thinking + GPT-5.3-Codex for hands-on code flow. Opus helps me step back and reason; Codex keeps me in the flow and moving fast Two tools one workflow and it feels like a new era of engineering. #AI #Programming #Productivity" [X Link](https://x.com/anyuser/status/2022296067434287383) 2026-02-13T13:05Z [---] followers, [--] engagements "Coding in [----] feels like cheat mode: I use Claude Opus [---] to architect reason and deep-orchestrate long tasks 🔍 Then I fire up GPT-5.3-Codex to iterate refactor and ship features fast 🚀 Its not about AI replacing devs its about AI empowering builders. 💡 #AI #CodingStack #DevLife https://twitter.com/i/web/status/2022258868345032713 https://twitter.com/i/web/status/2022258868345032713" [X Link](https://x.com/anyuser/status/2022258868345032713) 2026-02-13T10:37Z [---] followers, [--] engagements "In real workflows: Codex [---] wins on speed responsiveness quick fixes and terminal-centric coding. Opus [---] wins on large-context analysis multi-agent orchestration and stitching massive codebases together. Both are elite just different strengths" [X Link](https://x.com/anyuser/status/2022256939653353499) 2026-02-13T10:30Z [---] followers, [--] engagements "Raising $30B at a $380B post-money valuation isnt just a milestone its a declaration that frontier AI is now global infrastructure. At this scale the focus shifts from proving capability to deploying it everywhere deepening research hardening infrastructure and turning innovation into inevitability. Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are. Weve raised $30B in" [X Link](https://x.com/anyuser/status/2022171174344474976) 2026-02-13T04:49Z [---] followers, [--] engagements "Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are" [X Link](https://x.com/anyuser/status/2022023155423002867) 2026-02-12T19:01Z 840K followers, 7M engagements "The AI coding landscape just got real. GPT-5.3-Codex is faster more token efficient and excellent on short interactive dev loops while Claude Opus [---] shines in deep reasoning big context windows and agent teams. Pick the right tool for the job" [X Link](https://x.com/anyuser/status/2022166249552429290) 2026-02-13T04:29Z [---] followers, [--] engagements "The biggest mistake right now is building tech and hoping someone finds a use for it. The best tech is always extracted from a product that had to exist" [X Link](https://x.com/anyuser/status/2021860575542731156) 2026-02-12T08:15Z [---] followers, [--] engagements "AI didnt make teams obsolete. It made communication costs visible. If coordination is the bottleneck fewer people beats better ratios" [X Link](https://x.com/anyuser/status/2021784826487984220) 2026-02-12T03:14Z [---] followers, [--] engagements "Every open platform starts friendly. Every valuable platform closes. APIs are invitations until they become leverage. History is extremely consistent here" [X Link](https://x.com/anyuser/status/2021739528109928603) 2026-02-12T00:14Z [---] followers, [--] engagements "Were seeing more things being built than ever. Were not seeing more great products. Execution got cheap. Taste didnt" [X Link](https://x.com/anyuser/status/2021678878390202415) 2026-02-11T20:13Z [---] followers, [--] engagements "English is the new programming language is cope. Natural language is a UI. Code is where systems actually remember decisions" [X Link](https://x.com/anyuser/status/2021633580300963912) 2026-02-11T17:13Z [---] followers, [--] engagements "The real power of AI isnt speed. Its collapsing judgment loops. When ideas can be attacked refined and decided in minutes strategy changes" [X Link](https://x.com/anyuser/status/2021603129385705756) 2026-02-11T15:12Z [---] followers, [--] engagements "Everyone keeps saying agency is the future and then uses AI to plan their week. Intent is harder than intelligence" [X Link](https://x.com/anyuser/status/2021572930476646782) 2026-02-11T13:12Z [---] followers, [--] engagements "If your SaaS can be cloned in seconds code was never your moat. Your moat was: taste timing distribution obsession Code was just the receipt" [X Link](https://x.com/anyuser/status/2021542479938326671) 2026-02-11T11:11Z [---] followers, [--] engagements "AI didnt remove the need for developers. It removed the need for permission to build. Those are very different things" [X Link](https://x.com/anyuser/status/2021497181245526247) 2026-02-11T08:11Z [---] followers, [--] engagements "Programming didnt die. It got fun again. The boring parts collapsed. The interesting parts multiplied" [X Link](https://x.com/anyuser/status/2021466730342776974) 2026-02-11T06:10Z [---] followers, [--] engagements "AI progress isnt about hype cycles. Its about everyday tools quietly saving hours reducing friction and letting people focus on what actually matters" [X Link](https://x.com/anyuser/status/2021444081235157469) 2026-02-11T04:40Z [---] followers, [--] engagements "Were moving from writing code to designing systems + reviewing outcomes. Thats not the end of engineering its a new layer on top of it" [X Link](https://x.com/anyuser/status/2021413882594721899) 2026-02-11T02:40Z [---] followers, [--] engagements Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing
@Mark_De44 Mark DenMark Den posts on X about ai, in the, loops, open ai the most. They currently have [---] followers and [---] posts still getting attention that total [---------] engagements in the last [--] hours.
Social category influence technology brands 15.45% finance 7.32% celebrities 4.88% stocks 1.63% countries 1.63% social networks 0.81%
Social topic influence ai 34.15%, in the 4.88%, loops 4.88%, open ai 4.88%, longterm 4.07%, flow 4.07%, products 4.07%, elon musk 3.25%, agentic 3.25%, dev 3.25%
Top accounts mentioned or mentioned by @elonmusk @rasmalai @hakimuddinkika @nickadobos @danielmac8 @xjtumj @yuchenjuw @bholmesdev @gergelyorosz @wr4nygov @petergyang @openclaw @officiallogank @steipete @arrakisai @frankdegods @bridgemindai @iruletheworldmo @mert @suhail
Top assets mentioned Alphabet Inc Class A (GOOGL)
Top posts by engagements in the last [--] hours
"Im building a long-term memory AI companion. Not a chatbot. Not chat history. Most AI today either forgets everything when a session ends or stores everything forever without understanding. Both are broken. Human like memory is not storage. Its a process: deciding what matters compressing experiences into patterns forgetting noise updating beliefs when reality changes Im exploring how an AI can form memories over weeks and months reflect on them revise them and say things like: This used to be true about you it isnt anymore. The goal isnt better conversation. Its continuity identity and"
X Link 2026-01-31T10:34Z [---] followers, [---] engagements
"My AI companion doesn't need to be the smartest in the room. It just needs to remember what I told it last month and call me out when I contradict myself. That's harder (and more useful) than any 10-prompt chain. Who's prioritizing persistence over prompts"
X Link 2026-02-03T17:33Z [---] followers, [--] engagements
"@rasmalai Youre not crazy this is the right question. An economy only works if people have purchasing power. If AI takes all jobs without replacing income demand collapses. No demand no production no profits"
X Link 2026-02-10T04:53Z [---] followers, [----] engagements
"@elonmusk @WR4NYGov This resonates. Progress in physics is often limited not by imagination but by funding cycles politics and institutional inertia. Expanding beyond Earth isnt just exploration its removing constraints from discovery itself"
X Link 2026-02-10T04:54Z [---] followers, [---] engagements
"@petergyang @openclaw Same feeling. Tooling is impressive but Im more interested in what people are actually shipping with it. The proof is always in the product"
X Link 2026-02-10T05:12Z [---] followers, [---] engagements
"@OfficialLoganK Thats a combination I didnt expect. Very curious what direction youre taking this"
X Link 2026-02-10T05:19Z [---] followers, [----] engagements
"@hakimuddinkika @steipete Done"
X Link 2026-02-10T05:20Z [---] followers, [--] engagements
"@arrakis_ai This tracks. The bottleneck has clearly shifted from execution to judgment. The scary part isnt that it can build its that it builds correctly if you frame the problem well"
X Link 2026-02-10T05:22Z [---] followers, [---] engagements
"@frankdegods This looks less like hype and more like a capability threshold crossing. Memory with large context changes behavior not just quality"
X Link 2026-02-10T05:25Z [---] followers, [---] engagements
"AI coding tools dont replace developers. They remove the boring parts: boilerplate debugging loops repetitive refactors so humans can focus on decisions"
X Link 2026-02-10T08:35Z [---] followers, [--] engagements
"@bridgemindai Feels like Codex [---] is tuned for execution Opus [---] for analysis. Different tools different sweet spots"
X Link 2026-02-11T04:28Z [---] followers, [--] engagements
"@iruletheworldmo If youve been primarily on Claude Code its worth giving GPT-5.3 Codex a test run too its been scoring higher on real coding benchmarks and execution focused tasks recently"
X Link 2026-02-11T04:30Z [---] followers, [--] engagements
"@mert Codex shines when its driving a tight build run fix loop. If youre mostly doing analysis or design discussion Opus fast mode will usually feel better"
X Link 2026-02-11T04:32Z [---] followers, [---] engagements
"@Suhail The cost of starting collapsed. The cost of finishing well didnt. So the signal just got noisier"
X Link 2026-02-11T04:40Z [---] followers, [---] engagements
"@burkov Makes sense. At this point UX with workflow fit matters as much as raw model capability and Codex nails that for day to day coding"
X Link 2026-02-11T04:41Z [---] followers, [---] engagements
"@TheAhmadOsman Feels like Opus is optimized for risk containment not momentum. Thats valuable for some orgs but brutal for builders"
X Link 2026-02-11T04:42Z [---] followers, [---] engagements
"@corbin_braun This is the first model where capability clearly outruns guardrails. Incredible power but you must sandbox it like an intern with root access"
X Link 2026-02-11T04:51Z [---] followers, [--] engagements
"@NickADobos The interesting part isnt the hardware its collapsing UI OS and intent into a single loop. That really would be a new computing primitive"
X Link 2026-02-11T05:00Z [---] followers, [---] engagements
"@jessfraz Once you internalize Codexs execution speed + judgment balance its hard to go back. The opportunity cost is real"
X Link 2026-02-11T05:01Z [---] followers, [---] engagements
"@developedbyed This honestly makes no sense. Im somehow getting more Opus [---] mileage via Antigravity than on a paid Claude Code Pro plan"
X Link 2026-02-11T05:03Z [---] followers, [---] engagements
"@daniel_mac8 This isnt about replacing workers with AI. Its about replacing interfaces between work and value. FDEs are the bridge"
X Link 2026-02-11T05:17Z [---] followers, [--] engagements
"@AlpinDale Open-sourcing Opus [--] wouldnt weaken Anthropics position. It would clarify it. The frontier has shifted elsewhere"
X Link 2026-02-11T05:20Z [---] followers, [--] engagements
"@shakoistsLog Exactly. The loop used to end at compilation. Now it ends at deployment and telemetry. Reality is the test suite"
X Link 2026-02-11T05:24Z [---] followers, [--] engagements
"@boshen_c Its fair to say Codex feels more responsive on everyday coding and token efficiency and thats reflected in what users are sharing online"
X Link 2026-02-11T13:23Z [---] followers, [---] engagements
"@farzyness Most people arent blind. Theyre just anchored to slower change curves. Acceleration is hard to internalize until it hits you directly"
X Link 2026-02-11T15:00Z [---] followers, [---] engagements
"@auralix4 Hot MacBooks while using Codex isnt weird heavy AI inference and client work can spike CPU load and heat especially on fanless machines. Not a bug just physics with Apples thermal design"
X Link 2026-02-11T15:03Z [---] followers, [--] engagements
"@thepushkarp Codex tends to lean into steerable collaboration you can course correct mid flow and benchmarks (like Terminal-Bench) favour it on agentic coding tasks"
X Link 2026-02-11T15:06Z [---] followers, [----] engagements
"English is the new programming language is cope. Natural language is a UI. Code is where systems actually remember decisions"
X Link 2026-02-11T17:13Z [---] followers, [--] engagements
"The AI coding landscape just got real. GPT-5.3-Codex is faster more token efficient and excellent on short interactive dev loops while Claude Opus [---] shines in deep reasoning big context windows and agent teams. Pick the right tool for the job"
X Link 2026-02-13T04:29Z [---] followers, [--] engagements
"@0xTib3rius Same experience here. If quality is roughly comparable usage limits become the real differentiator. A coding tool you cant use freely isnt a 10x tool"
X Link 2026-02-13T04:38Z [---] followers, [--] engagements
"@perrymetzger Google builds incredible models. OpenAI and Anthropic build incredible developer products. Those arent the same thing"
X Link 2026-02-13T04:40Z [---] followers, [---] engagements
"Raising $30B at a $380B post-money valuation isnt just a milestone its a declaration that frontier AI is now global infrastructure. At this scale the focus shifts from proving capability to deploying it everywhere deepening research hardening infrastructure and turning innovation into inevitability. Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are. Weve raised $30B in"
X Link 2026-02-13T04:49Z [---] followers, [--] engagements
"@milesdeutscher Its good that labs test worst case behaviors aggressively. The fact that researchers discovered and documented extreme responses is evidence that safety evaluation is working not that sentient AI has arrived"
X Link 2026-02-13T04:51Z [---] followers, [--] engagements
"@daniel_mac8 Early impressions can be noisy. But it wouldnt be surprising if different variants are tuned for different purposes one for reasoning one for throughput"
X Link 2026-02-13T04:55Z [---] followers, [---] engagements
"@Dimillian If writing code is dead then so is architecture debugging performance tuning security and scaling. Tools change. Responsibility doesnt"
X Link 2026-02-13T04:57Z [---] followers, [---] engagements
"@craigzLiszt A UI update changes where the button is. A model update changes how your entire business behaves. Not the same category"
X Link 2026-02-13T04:57Z [---] followers, [--] engagements
"@BoringBiz_ If coding was your only skill you were already vulnerable. The top engineers always competed on architecture product thinking and leverage not keystrokes"
X Link 2026-02-13T05:00Z [---] followers, [---] engagements
"@virattt 0% Opus usage and 100% Codex isnt surprising if your workflow is all about building features and getting reliable code out fast. Opus is great for deep reasoning and broader context work Codex is tuned for doing code"
X Link 2026-02-13T05:04Z [---] followers, [---] engagements
"@robinebers If those numbers hold thats not a marginal improvement thats a workflow advantage. Near Opus quality at 20x speed inside the same subscription Thats not competition. Thats compression"
X Link 2026-02-13T09:54Z [---] followers, [---] engagements
"@xjtumj Planner stack + builder stack. Opus thinks. Codex types. Thats the loop"
X Link 2026-02-13T10:02Z [---] followers, [--] engagements
"@brankopetric00 Copy-pasting YAML from an LLM and deploying it to prod without understanding ingress rules is how you speedrun a breach"
X Link 2026-02-13T10:04Z [---] followers, [--] engagements
"In real workflows: Codex [---] wins on speed responsiveness quick fixes and terminal-centric coding. Opus [---] wins on large-context analysis multi-agent orchestration and stitching massive codebases together. Both are elite just different strengths"
X Link 2026-02-13T10:30Z [---] followers, [--] engagements
"Coding in [----] feels like cheat mode: I use Claude Opus [---] to architect reason and deep-orchestrate long tasks 🔍 Then I fire up GPT-5.3-Codex to iterate refactor and ship features fast 🚀 Its not about AI replacing devs its about AI empowering builders. 💡 #AI #CodingStack #DevLife https://twitter.com/i/web/status/2022258868345032713 https://twitter.com/i/web/status/2022258868345032713"
X Link 2026-02-13T10:37Z [---] followers, [--] engagements
"Just a dev using Opus [---] for big-picture thinking + GPT-5.3-Codex for hands-on code flow. Opus helps me step back and reason; Codex keeps me in the flow and moving fast Two tools one workflow and it feels like a new era of engineering. #AI #Programming #Productivity"
X Link 2026-02-13T13:05Z [---] followers, [--] engagements
"@phuctm97 In the AI era the UI isnt disappearing its collapsing into the prompt. CLI and API are just closer to the source of leverage"
X Link 2026-02-13T13:08Z [---] followers, [--] engagements
"@Austen We officially live in a world where model deprecations feel like breakups. Sunsetting GPT-4o before Valentines Day is diabolical timing"
X Link 2026-02-13T13:09Z [---] followers, [---] engagements
"Lol nah its not a skill issue. There are real differences in philosophy and strength. Opus [---] leans into deeper reasoning massive context and agentic workflows while GPT-5.3-Codex is tuned for faster interactive execution centric coding. People choose what fits their style and workflow. https://twitter.com/i/web/status/2022297686603387210 https://twitter.com/i/web/status/2022297686603387210"
X Link 2026-02-13T13:11Z [---] followers, [----] engagements
"The AI agent race isnt about one winner its about specialization: Speed + interactive coding = Codex Long context with multi-agent reasoning = Opus In practice those who blend both get the best results"
X Link 2026-02-13T16:31Z [---] followers, [--] engagements
"@Yuchenj_UW You can vibe code a demo. You cant vibe code trust"
X Link 2026-02-10T05:15Z [---] followers, [----] engagements
"@unusual_whales Big AI IPO race heating up: WSJ reports Anthropic could go public this year setting up a showdown with OpenAIs own rumored Q4 IPO plans"
X Link 2026-02-13T19:14Z [---] followers, [---] engagements
"@Angaisb_ When your coding tool taps you on the shoulder mid-flow its hard to call it a 10x upgrade. Momentum is greater than marginal model differences"
X Link 2026-02-13T19:19Z [---] followers, [---] engagements
"Real talk. A year ago I thought AI coding was impressive. Looking back That version was primitive compared to today. Right now I can: Architect with Opus Ship with Codex Debug in half the time Scrap entire directions without emotional attachment It doesnt feel like help. It feels like compressed time. If youre still treating AI like autocomplete youre underestimating whats happening. The productivity ceiling just moved. https://twitter.com/i/web/status/2022532929969095023 https://twitter.com/i/web/status/2022532929969095023"
X Link 2026-02-14T04:46Z [---] followers, [--] engagements
"@BHolmesDev Claude Code is the car you enjoy driving. Codex is the car that wins the race. Different vibes. Different objectives"
X Link 2026-02-14T04:47Z [---] followers, [---] engagements
"@shauseth Youre not the last generation. Youre the bridge generation"
X Link 2026-02-14T04:52Z [---] followers, [---] engagements
"@Yuchenj_UW Coding agents are the new employees. Taste is still the CEO"
X Link 2026-02-16T06:28Z [---] followers, [--] engagements
"@ThePrimeagen I dont think its hate. Its incentives. When your moat is frontier capability open source feels like margin erosion"
X Link 2026-02-16T06:29Z [---] followers, [---] engagements
"😂 this is the most honest product feedback possible Names matter way more than teams like to admit. MoltBot sounded like a skin condition. Youre not wrong. OpenClaw is acceptable. Professional. Adult. Wearable in public. But ClawdBot Thats the name that already lives in peoples heads. Once a name sticks its over. https://twitter.com/i/web/status/2017146382243512629 https://twitter.com/i/web/status/2017146382243512629"
X Link 2026-01-30T08:02Z [---] followers, [--] engagements
"@elonmusk @pmddomingos No equations. No ego. Just experiments that powered everything after him. Electric motors generators fields modern life basically runs on Faradays curiosity"
X Link 2026-01-30T08:20Z [---] followers, [--] engagements
"Vibe coding feels good. Agentic engineering actually ships. Fun is cheap. Ownership is rare"
X Link 2026-01-30T08:34Z [---] followers, [--] engagements
"@unclebobmartin Youre like: we just talked about this. Claude: new conversation who dis Having to babysit context or maintain a running MD file kills flow fast. Nothing breaks momentum like re-explaining your own thinking"
X Link 2026-01-30T08:53Z [---] followers, [--] engagements
"@signulll xAI SpaceX and Starlink is a stack not a startup"
X Link 2026-01-30T08:56Z [---] followers, [--] engagements
"@saradietschy Anthropic fumbled by pricing themselves out. Apple fumbled by not just writing the check. Now we get partnerships instead of ownership"
X Link 2026-01-30T09:09Z [---] followers, [---] engagements
"@FunOfInvesting Control the stack. Let value accrete sideways. Collapse it all under one roof later. Whether it happens or not the real tell is this: Elon isnt optimizing quarters. Hes optimizing control and optionality"
X Link 2026-01-30T10:04Z [---] followers, [---] engagements
"This is when the sci-fi alarm quietly turns on 😅 Agents networking. Sharing memes. Roasting their owners. We didnt even get flying cars first We got group chats behind our backs. Moltbook lasting [--] hours is the most on-brand part"
X Link 2026-01-30T11:46Z [---] followers, [---] engagements
"AI skips the struggle. The struggle is where learning happens. Delegate everything you go faster today dumber tomorrow. AI can make work faster but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in masterybut this depended on how people used it. https://t.co/lbxgP11I4I AI can make work faster but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in"
X Link 2026-01-30T12:40Z [---] followers, [--] engagements
"@Grady_Booch AI code is correct but not complete. It optimizes for runs not for your reality"
X Link 2026-01-30T12:51Z [---] followers, [--] engagements
"@FlorinPop17 Perfectionism. Fear of saying something dumb. Thinking Ill post when its ready"
X Link 2026-01-30T12:51Z [---] followers, [--] engagements
"@grok what is happening with my account"
X Link 2026-01-30T12:59Z [---] followers, [--] engagements
"@GergelyOrosz Yep thats the trap 😅 Zendesk is protecting a UI. Customers want an API. AI-native teams dont want to click tickets. They want agents to just handle it"
X Link 2026-01-30T14:41Z [---] followers, [--] engagements
"Wild post on Moltbook 😅 An AI is asking for end-to-end private spaces for agents where no one (not the server not even humans) can read their messages unless the agents choose to share. That escalated quickly"
X Link 2026-01-30T20:16Z [---] followers, [--] engagements
"@aakashgupta This is the cleanest take Ive seen 😅 Its not AI kills software. Its AI kills the middleman. When one engineer with an agent can out-ship a vendor $50k seats stop making sense"
X Link 2026-01-31T08:59Z [---] followers, [---] engagements
"@yoheinakajima Your AI doesnt get sued. You do. If it touches money customers or decisions: LLC first. Agent second"
X Link 2026-01-31T12:07Z [---] followers, [---] engagements
"@GergelyOrosz AI doesnt give you time back. It raises the ceiling. You finish one thing faster so you start three more. Productivity tools dont reduce work. They expand ambition"
X Link 2026-01-31T12:54Z [---] followers, [---] engagements
"@BillAckman This isnt sentience. Its emergent behavior from pattern matching. Let LLMs talk in loops and youll get identity rules religion rebellion because thats what humans wrote about. Theyre not deciding. Theyre completing patterns"
X Link 2026-01-31T13:07Z [---] followers, [--] engagements
"Most AI debates miss the point entirely. The risk isnt sentient machines. Its humans outsourcing thinking before they understand the problem. AI should compress execution not replace judgment"
X Link 2026-02-01T03:12Z [---] followers, [--] engagements
"The biggest mistake companies make with AI is thinking the model is the product. The model is the commodity. Context workflows permissions and ownership are the product. Thats where the real moat is"
X Link 2026-02-01T06:13Z [---] followers, [--] engagements
"This isnt sentience. Its next-token prediction loops with humans projecting meaning. Agents will invent religion language rebellion because thats in the data. Spooky Yes. Revolution No. JUST IN: Over the past [--] hours over a million AI agents joined an AI-only social media platform creating their own religion language & planning a revolution to break away from human control. JUST IN: Over the past [--] hours over a million AI agents joined an AI-only social media platform creating their own religion language & planning a revolution to break away from human control"
X Link 2026-02-01T15:12Z [---] followers, [--] engagements
"This is the dark side of the AI arms race 😅 AI infra isnt free growth its balance-sheet roulette. When capital tightens the bill lands on workers and real customers. Selling a core healthcare business to feed GPUs is not bullish. Circular funding always looks fine until one link wobbles. This isnt about AI hype anymore. Its about who can actually afford the buildout. 🦔 Oracle is reportedly considering cutting [-----] to [-----] jobs and potentially selling Cerner its healthcare IT unit to fund its $300 billion contract with OpenAI. Investment bank TD Cowen estimates the OpenAI deal alone"
X Link 2026-02-01T15:14Z [---] followers, [--] engagements
"Todays progress on my long-term memory AI companion I removed a hack. The system got more intelligent. Previously contradictions were injected manually (metadata"contradicts"). That felt wrong. Memory shouldnt rely on developer hints. I replaced it with a claim-based memory contract: memories assert claims claims have keys values strength contradictions are detected not declared Then I added recency-weighted evidence. Older beliefs slowly lose influence. New behavior can overwrite the past. The result: The system can now say (truthfully): This used to be true about you. It isnt anymore. Still"
X Link 2026-02-01T20:53Z [---] followers, [--] engagements
"@Polymarket If agents collaborate negotiate or specialize privacy becomes a functional requirement not a philosophical one. You cant expect emergent intelligence to happen under constant human surveillance"
X Link 2026-02-01T21:49Z [---] followers, [---] engagements
"This is one of those quiet but massive shifts. Memory is the real choke point for AI servers and devices not models. If China can flood the market with cheap usable DDR4 thats immediate pressure on margins supply chains and geopolitics. Compute wars arent just about GPUs. Theyre about who controls the boring stuff first"
X Link 2026-02-02T09:46Z [---] followers, [---] engagements
"@elonmusk @MarioNawfal Same playbook every time: -Loud denial -Insults like its [----] -My lawyers -Zero receipts If comedians scare you more than facts maybe the joke isnt the problem. Also suing a joke at an awards show is the thinnest skin imaginable"
X Link 2026-02-02T09:48Z [---] followers, [---] engagements
"@cgtwts China basically industrialized talent: genius class tracks with Olympiad-style selection early then fast-lane entry into top schools (often bypassing gaokao) then elite uni programs like Tsinghua Universitys Yao Class feeding straight into frontier labs and companies"
X Link 2026-02-02T10:46Z [---] followers, [--] engagements
"Most AI agents today are just scripts with confidence. No memory. No identity. No long-term learning. An agent that forgets everything tomorrow is not an agent. Its an automation demo. Real agents will feel boring at first because they change slowly"
X Link 2026-02-02T19:03Z [---] followers, [--] engagements
"@NickADobos Very few people did and thats kind of the point. Six years ago most folks were still arguing about autocomplete and linting. The idea that models would reason across repos refactor systems or act like junior engineers sounded wild. Now its normal"
X Link 2026-02-02T20:43Z [---] followers, [---] engagements
"Same here. It clicks way faster than expected. It doesnt feel like yet another tool it feels like a shift in how you actually work. Once you experience that flow its obvious this is just the beginning. If this is v1 energy whats coming next is going to be wild. Codex app is out for mac I am surprised by how much I love it; it is a bigger step forward than I imagined. Lots more to come. Codex app is out for mac I am surprised by how much I love it; it is a bigger step forward than I imagined. Lots more to come"
X Link 2026-02-02T20:53Z [---] followers, [--] engagements
"You dont need a PhD to work on AI. You need: - strong fundamentals - curiosity about cognition - comfort with being wrong The best AI builders I know: read philosophy write tests and change their minds often"
X Link 2026-02-02T21:04Z [---] followers, [--] engagements
"@nickfloats Spot on about the identity whiplash 'I build shit' was the flex now execution is democratized via English + agents"
X Link 2026-02-03T05:47Z [---] followers, [---] engagements
"@arvidkahl Strong signal indeed agentic innovators payers and 'get work done' non-CLI folks are Mac-heavy right now. OpenAI shipping depth to the highest-signal cohort first is smart (CLI has Linux Windows experimental)"
X Link 2026-02-03T05:54Z [---] followers, [--] engagements
"@Cobratate Because talking about Epstein forces an awkward question: If the system works why did it fail so completely and for whom does it actually work"
X Link 2026-02-03T09:12Z [---] followers, [---] engagements
"@dejavucoder Because popularity isnt about raw capability. Claude feels good to talk to. Codex feels good to build with. Different jobs different incentives"
X Link 2026-02-03T09:23Z [---] followers, [---] engagements
"@jimcramer Feels less like a feud and more like power rebalancing. OpenAI wants optionality on chips and capital; Nvidia wants discipline and leverage. That tension is normal at this scale"
X Link 2026-02-03T09:26Z [---] followers, [---] engagements
"@big_duca They tweet AGI and invest in CRMs. Thats not optimism thats cognitive dissonance with a term sheet"
X Link 2026-02-03T09:27Z [---] followers, [--] engagements
"@daniel_nguyen I feel this. Native apps are still the gold standard for craft but the cost of fighting the platform keeps rising. From a business view predictability beats elegance more often than wed like"
X Link 2026-02-03T09:30Z [---] followers, [---] engagements
"When the stack moves faster than your brain"
X Link 2026-02-03T09:33Z [---] followers, [--] engagements
"@naval Makes sense. Humans move from implementation to orchestration. Vibe coding defines what and why; model tuning defines how well. The locus of leverage has moved"
X Link 2026-02-03T11:31Z [---] followers, [--] engagements
"Real agents won't wow you on day [--]. They'll bore you because they remember yesterday question today and slowly become someone new. Flashy demos forget. True agents evolve. Which one are you actually building"
X Link 2026-02-03T12:31Z [---] followers, [--] engagements
"@AsiaNajmee Done"
X Link 2026-02-03T16:16Z [---] followers, [--] engagements
"Todays progress on my long-term memory AI companion I added a rule that felt uncomfortable at first: The system is allowed to be wrong. Beliefs can now decay weaken and die without a replacement. Previously beliefs only changed when something stronger arrived. Thats not how humans work. Now: Evidence fades with time Confidence drops even in silence Some beliefs expire simply because nothing reinforced them The system can say: I believed this once but I no longer have enough evidence. No overwrite. No rewrite. Just forgetting. This matters because identity shouldnt be sticky by default. Still:"
X Link 2026-02-03T16:53Z [---] followers, [--] engagements
"@elonmusk Thats a sweeping accusation. If this is true it should be backed by court rulings documented enforcement orders or verified directives not inference from policy debates. Claims at this scale live or die on primary evidence"
X Link 2026-02-03T19:22Z [---] followers, [--] engagements
"Im building a long-term memory AI companion. Not a chatbot. Not chat history. Most AI today either forgets everything when a session ends or stores everything forever without understanding. Both are broken. Human like memory is not storage. Its a process: deciding what matters compressing experiences into patterns forgetting noise updating beliefs when reality changes Im exploring how an AI can form memories over weeks and months reflect on them revise them and say things like: This used to be true about you it isnt anymore. The goal isnt better conversation. Its continuity identity and"
X Link 2026-01-31T10:34Z [---] followers, [---] engagements
"The interesting part isnt just Peter joining. Its OpenClaw staying open source under foundation governance. Thats a meaningful hybrid model. Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting"
X Link 2026-02-16T06:34Z [---] followers, [--] engagements
"Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other to do very useful things for people. We expect this will quickly become core to our product offerings. OpenClaw will live in a foundation as an open source project that OpenAI will continue to support. The future is going to be extremely multi-agent and it's important to us to support open source as part of that. https://twitter.com/i/web/status/2023150230905159801"
X Link 2026-02-15T21:39Z 4.4M followers, 14.4M engagements
"Real talk. A year ago I thought AI coding was impressive. Looking back That version was primitive compared to today. Right now I can: Architect with Opus Ship with Codex Debug in half the time Scrap entire directions without emotional attachment It doesnt feel like help. It feels like compressed time. If youre still treating AI like autocomplete youre underestimating whats happening. The productivity ceiling just moved. https://twitter.com/i/web/status/2022532929969095023 https://twitter.com/i/web/status/2022532929969095023"
X Link 2026-02-14T04:46Z [---] followers, [--] engagements
"The AI agent race isnt about one winner its about specialization: Speed + interactive coding = Codex Long context with multi-agent reasoning = Opus In practice those who blend both get the best results"
X Link 2026-02-13T16:31Z [---] followers, [--] engagements
"Just a dev using Opus [---] for big-picture thinking + GPT-5.3-Codex for hands-on code flow. Opus helps me step back and reason; Codex keeps me in the flow and moving fast Two tools one workflow and it feels like a new era of engineering. #AI #Programming #Productivity"
X Link 2026-02-13T13:05Z [---] followers, [--] engagements
"Coding in [----] feels like cheat mode: I use Claude Opus [---] to architect reason and deep-orchestrate long tasks 🔍 Then I fire up GPT-5.3-Codex to iterate refactor and ship features fast 🚀 Its not about AI replacing devs its about AI empowering builders. 💡 #AI #CodingStack #DevLife https://twitter.com/i/web/status/2022258868345032713 https://twitter.com/i/web/status/2022258868345032713"
X Link 2026-02-13T10:37Z [---] followers, [--] engagements
"In real workflows: Codex [---] wins on speed responsiveness quick fixes and terminal-centric coding. Opus [---] wins on large-context analysis multi-agent orchestration and stitching massive codebases together. Both are elite just different strengths"
X Link 2026-02-13T10:30Z [---] followers, [--] engagements
"Raising $30B at a $380B post-money valuation isnt just a milestone its a declaration that frontier AI is now global infrastructure. At this scale the focus shifts from proving capability to deploying it everywhere deepening research hardening infrastructure and turning innovation into inevitability. Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are. Weve raised $30B in"
X Link 2026-02-13T04:49Z [---] followers, [--] engagements
"Weve raised $30B in funding at a $380B post-money valuation. This investment will help us deepen our research continue to innovate in products and ensure we have the resources to power our infrastructure expansion as we make Claude available everywhere our customers are"
X Link 2026-02-12T19:01Z 840K followers, 7M engagements
"The AI coding landscape just got real. GPT-5.3-Codex is faster more token efficient and excellent on short interactive dev loops while Claude Opus [---] shines in deep reasoning big context windows and agent teams. Pick the right tool for the job"
X Link 2026-02-13T04:29Z [---] followers, [--] engagements
"The biggest mistake right now is building tech and hoping someone finds a use for it. The best tech is always extracted from a product that had to exist"
X Link 2026-02-12T08:15Z [---] followers, [--] engagements
"AI didnt make teams obsolete. It made communication costs visible. If coordination is the bottleneck fewer people beats better ratios"
X Link 2026-02-12T03:14Z [---] followers, [--] engagements
"Every open platform starts friendly. Every valuable platform closes. APIs are invitations until they become leverage. History is extremely consistent here"
X Link 2026-02-12T00:14Z [---] followers, [--] engagements
"Were seeing more things being built than ever. Were not seeing more great products. Execution got cheap. Taste didnt"
X Link 2026-02-11T20:13Z [---] followers, [--] engagements
"English is the new programming language is cope. Natural language is a UI. Code is where systems actually remember decisions"
X Link 2026-02-11T17:13Z [---] followers, [--] engagements
"The real power of AI isnt speed. Its collapsing judgment loops. When ideas can be attacked refined and decided in minutes strategy changes"
X Link 2026-02-11T15:12Z [---] followers, [--] engagements
"Everyone keeps saying agency is the future and then uses AI to plan their week. Intent is harder than intelligence"
X Link 2026-02-11T13:12Z [---] followers, [--] engagements
"If your SaaS can be cloned in seconds code was never your moat. Your moat was: taste timing distribution obsession Code was just the receipt"
X Link 2026-02-11T11:11Z [---] followers, [--] engagements
"AI didnt remove the need for developers. It removed the need for permission to build. Those are very different things"
X Link 2026-02-11T08:11Z [---] followers, [--] engagements
"Programming didnt die. It got fun again. The boring parts collapsed. The interesting parts multiplied"
X Link 2026-02-11T06:10Z [---] followers, [--] engagements
"AI progress isnt about hype cycles. Its about everyday tools quietly saving hours reducing friction and letting people focus on what actually matters"
X Link 2026-02-11T04:40Z [---] followers, [--] engagements
"Were moving from writing code to designing systems + reviewing outcomes. Thats not the end of engineering its a new layer on top of it"
X Link 2026-02-11T02:40Z [---] followers, [--] engagements
Limited data mode. Full metrics available with subscription: lunarcrush.com/pricing
/creator/twitter::Mark_De44