thank you, will send you my loom i sent my EA showing how incredibly beyond a disaster gusto is
Chronological feed of everything captured from swyx.
thank you, will send you my loom i sent my EA showing how incredibly beyond a disaster gusto is
i was going to send him my loom showing him gustos bugs and loom loomed on me
can you host this somewhere on a blog? (else happy to do it for u with full credit)
i'm just an advisor, cant take credit. team is doing great with devin 3.0 and devin desktop
try for nontech stuff too https://www.youtube.com/watch?v=zepu8Kk6FBQ
swyx proposes a Chrome extension to enhance all web image input boxes with three non-AI options: generating simple word-based text graphics, drawing via @tldraw integration, or using either to produce images matching required dimensions. This targets seamless augmentation of upload interfaces without relying on AI models. The request is directed at @devinai for implementation.
Codex provides superior AI-powered search capabilities within Slack compared to Slack's own AI search feature. This assessment comes from an hourly poll on swyx's X feed, highlighting Codex's edge in practical messaging app search tasks. Technical users may prefer Codex for enhanced retrieval accuracy in collaborative tools.
Whatcable, previously referenced in swyx's hourly X feed poll, now offers a dedicated Mac application for users. The open-source tool is hosted on GitHub at https://github.com/darrylmorley/whatcable. This enables seamless, native desktop access to X (Twitter) feeds without browser dependency.
Among frontier models, xAI's Grok 4.30 delivers the highest intelligence per dollar, surpassing open models like MiMo, Kimi, and DeepSeek. Codex has become a strict superset of ChatGPT, prompting its app uninstallation, and excels as a Slack AI search alternative. This shift emphasizes efficiency in AI tool selection.
Among frontier models, xAI's Grok 4.30 delivers the highest intelligence per dollar, outperforming open models like MiMo, Kimi, and DeepSeek. Data sourced from Artificial Analysis confirms this cost-efficiency edge. Users can verify independently by replicating the analysis.
Among frontier models, xAI's Grok 4.30 delivers the highest intelligence per dollar, surpassing open models like MiMo, Kimi, and DeepSeek. Data sourced from Artificial Analysis highlights this cost-efficiency edge. This milestone underscores shifting priorities toward model efficiency over raw capability.
A technical poll on swyx's X feed questions if AI models conduct image search during their "thinking" process and whether they can yet perform visual comparisons. This highlights ongoing curiosity in the AI community about multimodal reasoning capabilities. No definitive answers are provided in the content, framing it as an open hourly discussion point.
Swyx's X feed shows a clear upward trajectory worth monitoring hourly. Hardware advancements are positioned as the decisive factor shaping this path. The phrase "hardware is destiny" underscores determinism in tech evolution via physical infrastructure.
An hourly poll on swyx's X feed received a single visible response of "YESS," indicating positive user sentiment. This affirmative reaction suggests high engagement or approval from at least one participant in the poll. No additional responses or metrics are provided in the content.
X has unexpectedly reintroduced hourly polls in swyx's feed, a feature the author believed had been removed. This revival prompts calls for technical analysis comparing the initial implementation to the current version. Differences in architecture, performance, or user experience could reveal platform iteration strategies.
A user-conducted hourly poll on swyx's X feed resulted in a "sad" assessment. This single-word response indicates negative emotional reception to the feed's content. No additional metrics or context provided.
Coding agents are breaking out of their specialized domain to handle diverse non-coding tasks, marking 2024 as the year knowledge workers broadly adopt AGI-level tools. A tiny team leverages multiple agents like OpenClaw, Devin, and TownAI to operate @aidotengineer, serving 1M unique developers monthly for free across applications from CMS to niche rentals. This shift, amplified by upgrades like Codex for general computer work, urges broader experimentation in daily workflows.
Codex app's latest update adds /side command or menu option to spawn side chats within the current context, ideal for tangential questions during deep threads. The CLI version already includes this /side feature. Users are requesting similar shorthand like /btw for convenience.
Swyx publicly congratulates @ajambrosino on a significant product or software update. The update is highlighted for eliminating "clunky handoff/switching," indicating improved user flow without interruptions. Swyx seeks elaboration on this specific feature's meaning or reference.
Swyx runs hourly polls on X to capture real-time AI news sentiment, linking to fuller writeups like Latent Space's AI News: Agents for Everything Else. This method distills trending topics in AI development. It highlights a pivot from coding agents to broader "agents for everything else" paradigms.
Direct fetch of X post ID 2050076374082429248 by @swyx failed. Searches for recent @swyx posts yielded only ID 2052538105887150267 from May 7, 2026, indicating the target post is inaccessible. As of May 8, 2026, it is likely deleted, private, or does not exist; @swyx's profile describes a tech/AI influencer with 156k followers and affiliations including @dxtipshq, @cognition, and @temporalio.
Swyx's X feed includes an hourly poll mechanism noted by users. It also prominently features a "kind bot," suggesting automated friendly interactions. This setup indicates a focus on regular engagement tools in the feed.
Attempts to fetch X post ID 2050084459291324673 failed due to potential deletion, privacy settings, or tool limitations. Recent @swyx posts cover AI tools, podcasts, and Singapore food, but none correspond to the target ID. This highlights reliability issues in automated social media content retrieval for archival or analysis pipelines.
A long-term resident of Singapore reports observable annual enhancements in public infrastructure, including new MRT stations, HDBs, condos, and public spaces, alongside better restaurants. This consistent progress over half a decade, compared to experiences in 7 countries, leads to the assessment of Singapore as the best-run country on Earth. The post promotes an AI engineering event in Singapore with incentives like personal tours to top local eateries.
Notion pioneered custom agents integrated with its powerful editor, prompting OpenAI to launch similar "workspace agents" in ChatGPT for complex, team-based workflows. OpenAI lacks an equivalent editor, limiting full replication. Notion refined its AI through five rebuilds, as detailed in recent podcast and profile discussions.
Swyx expresses strong personal excitement for a potential GPT-5.5 launch on his birthday, saluting Sam Altman's post in an hourly poll context. This builds into a thread where he signals "its happening" by quoting OpenAI's "NS41" developer post. The posts indicate insider-like anticipation for an imminent OpenAI model release.
Image-2-Thinking is an agentic extension of the Image-2 model that employs search and Photoshop tools in a loop to generate complex images like QR codes, diagrams, logos, foods, and faces in tens of minutes. This mirrors Gemini Flash Vision's benchmark dominance through agentic image-to-text processing. The approach enables one-shot creation of intricate visuals by compositing and self-reviewing outputs.
Swyx is recognized as a singular leader propelling the AI engineering community forward through high-impact initiatives. He originated the AI Engineer concept, curates the acclaimed Latent Space podcast with top-tier guests, and maintains the influential No Need newsletter featuring "nothing major" updates. Community members highlight his exceptional work ethic and dedication to delivering timely AI news and events.
Swyx's hourly X feed poll responses hinge on meme involvement level. Simpler memes likely prompt quicker engagement, while involved ones demand more user consideration. This suggests tailoring content complexity to desired interaction depth in social feeds.
Modern LLMs, when prompted to generate image descriptions, spontaneously include specific cultural and local details like "third wave coffee" and "namma ooru" without explicit instruction, drawing directly from internalized training knowledge. This is evident in a DALL-E 3 era prompt depicting a Bengaluru girl in Cubbon Park, where such elements surprise users by their unprompted presence. Comparisons to 2023 images highlight the evolution or consistency of this capability in model outputs.
Swyx's X feed hourly poll concludes with a message wishing participants good luck. He explicitly invites them to reach out for any help. This reflects an open, supportive engagement style in his social media interactions.
In an hourly poll evaluating swyx's X feed, the user deems it "decent for 2023." This suggests the feed meets baseline quality expectations from that year. No further metrics or comparisons are provided.
Over ten years, the author transformed from career isolation and limited skills to building an influential AI engineering career, media empire, and rich personal life through relentless reinvention every 2-5 years. Echoing Gates' insight, short-term progress is overestimated while decade-scale achievements via self-belief—betting on personal visions despite external cynicism—prove transformative. Frequent pivots preserve neuroplasticity but demand defining and mastering one's own game for sustained satisfaction and impact.
Swyx's X feed functions as an hourly poll delivering timely updates. It includes additional references linked in footnotes to his "decade" page at swyx.io/decade. This setup provides a structured stream of insights for followers.
no no no tech bf medical gf this is known @NaderLikeLadder @nutlope
thanks man so nice to see you! definitely come to sf for aiewf!!