Dropbox status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Dropbox reports received over the last 24 hours by time of day. When the number of reports exceeds the baseline, represented by the red line, an outage is determined.

Most Reported Problems

The following are the most recent problems reported by Dropbox users through our website.

• Errors (60%)
• Sign in (20%)
• Website Down (20%)

Live Outage Map

The most recent Dropbox outage reports came from the following cities:

City	Problem Type	Report Time
Nottingham	Errors	4 days ago
Guayaquil	Website Down	4 days ago
Flumet	Errors	14 days ago
Irapuato	Errors	17 days ago
Bournemouth	Sign in	2 months ago
Paramaribo	Errors	3 months ago

Full Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Dropbox Issues Reports

Latest outage, problems and issue reports in social media:

• 
  Jens Kristensen (@JensKri20101733) reported 5 minutes ago

  Suggestion for @adamhfry, ChatGPT Consumer Product Lead: The new Google Drive integration made me wonder: why not take the same idea one step further and support local Windows files directly? No Google Drive. No OneDrive. Local storage, controlled ChatGPT access. A file should not have to be stored in the cloud at all. Cloud has done enough damage already. Cloud = Hell. There is an important distinction between cloud computing and cloud storage. Cloud computing means that ChatGPT performs the processing on OpenAI’s servers. Cloud storage means that documents are permanently stored with Google, Microsoft, Dropbox, or another cloud provider. The first may be a practical consequence of ChatGPT’s current architecture. The second is not. A much cleaner model would be: Local disk / NAS → temporary, explicitly authorized ChatGPT access → processing → result returned to local disk / NAS. For example, a user could right-click: G:\Project\Analysis.docx and select “Open with ChatGPT”. ChatGPT would then receive controlled access to that file — or perhaps to a user-authorized folder such as: G:\ChatGPT\ The user could specify whether access should be read-only or read/write. Original files could be protected, and output could automatically be written to a designated local \output folder. Then instructions could be as simple as: “Edit only section 17. Preserve all formatting.” “Analyze all documents in G:\ChatGPT\Project X.” “Compare these three PDFs.” “Edit Analysis.docx, but do not modify the original. Save the result in \output.” DOCX, XLSX and PPTX are not fundamentally unsuitable for this. They are largely ZIP containers containing XML files. The harder problem is preserving complex formatting, images, tables, comments, undo/versioning and accurate rendering. A local “ChatGPT File Bridge” for Windows could solve the access problem without requiring users to move their working files into Google Drive or OneDrive. The AI processing itself would not necessarily be local. Files, or the relevant parts of them, could still be transmitted to OpenAI for processing. But storage and file management could remain entirely local: local file → controlled ChatGPT access → processing → result back to local disk / NAS. No Google Drive. No OneDrive. No permanent cloud storage. No manual upload/download cycle. The user retains control over the file structure, filenames, versions, backups, applications and physical storage location. “Google Docs inside ChatGPT” is technically interesting. But “Local Files inside ChatGPT” would be the real game changer for the traditional Windows PC workflow. And OpenAI would not need to invent another file system. Windows already has a perfectly good one.

• 
  Rocinante (@wb9rms6gyz) reported 6 minutes ago

  @griffin_daly_ @AlexisCoe Which could well have been part of the arrangement. This piece of **** is actively live tweeting his daughters stuffy issues and shared a Dropbox with naked photos of her. He’s a lunatic. As someone who ripped Moreno a week ago…no lie, I think he’s legally constrained

• 
  Abhishek Singh (@0xlelouch_) reported 8 minutes ago

  System design question. How would you design Dropbox-style file sync with conflict handling? Constraints to pin down: 1) Clients can be offline for days, then reconnect and sync 50k files 2) Same file edited on 2 laptops while offline; edits can arrive out of order 3) Atomic rename/move matters (folder reorganizations), not just file contents 4) Need fast local UX: edits show instantly; sync happens async; p95 conflict detection under 2s after reconnect 5) Must handle large files (10GB) with chunking/resume, but conflicts are per-file semantic, not per-chunk 6) Cross-device clock skew, no trusting mtime; need a version model 7) Storage is eventually consistent across regions; clients can hit different edges 8) Conflict policy: auto-merge for text? duplicate files? keep both? how do you avoid conflict loops? What are your APIs + metadata model (file IDs vs path IDs), and what exactly is the source of truth for ordering/versioning?

• 
  Duaa 🤲🇮🇳 (@Duaa_1206) reported 9 minutes ago

  @himasoraakane Doing digital forensics on Dropbox error screens to investigate a suspended account is a whole new level of investigative work.

• 
  Mikemira (@storiesbyohama) reported 16 minutes ago

  Just imagine getting accepted into the most exclusive startup club on earth… Then being told you have two weeks to find a complete stranger to as your partner. That’s exactly what happened to Drew Houston in 2007. He had the idea for Dropbox. He had a rough demo. @ycombinator liked it. But @paulg was clear... Single founders rarely make it. You need a co-founder. Right now. Drew’s friends couldn’t join. Time was running out. What would you do? He put out the word. A mutual friend connected him to a quiet MIT student named Arash Ferdowsi. They had never met. They sat down in the student center. Talked for about two hours. About code. About the problem. About the future. At the end of that conversation Arash said yes. He dropped out of MIT the next week with only one semester left. Two weeks later they walked into the YC interview together. They got in. The rest is history: a company that became worth billions. It looked reckless. It felt like a shotgun wedding. Yet it worked because both were all-in from the first conversation. I’ve studied hundreds of startups that never made it past the idea stage. Most founders wait too long for the “perfect” partner. They overthink chemistry. They protect their equity. They miss the window. You can’t wait for certainty. Sometimes the right co-founder is the person willing to jump with you before the proof exists. The speed of that decision can be the difference between staying a solo dreamer and building something real. What would you risk in two weeks if the right person walked in?

• 
  Nathan (@arcane_bloom) reported 17 minutes ago

  He was OpenAI's first business hire in 2018. This week, after eight years, he walked out the door. > Brad Lightcap > studies economics and history at Duke, starts as a JP Morgan investment banking analyst > moves into strategic finance at Dropbox, then joins Y Combinator's Continuity Fund > meets Sam Altman through YC, gets pulled into a tiny nonprofit called OpenAI in 2018 as its first business hire > becomes CFO, then rises to COO, helps run the company through the ChatGPT launch and its climb to the most valuable startup on earth > moved off the COO title in April 2026 into a vague "special projects" role > in August, posts on X that he's leaving after eight years to "start something new" > his exit lands one month after product chief Fidji Simo also stepped down > walks away right as OpenAI preps a monster IPO on an $852 billion valuation

• 
  FarhanX_AI (@FarhanBuildsAI) reported 20 minutes ago

  SETTING #1: Startup Apps Nobody Asked For What it does: Every time you install a new program, it quietly adds itself to a list of apps that launch the second Windows boots, whether you use it daily or once a year. Why it kills performance: Your laptop isn't just starting Windows when you power it on. It's simultaneously launching Spotify, Steam, Adobe updaters, Dropbox, Zoom, and a dozen other programs all fighting for the same limited CPU and RAM at once. How to fix it: Ctrl + Shift + Esc to open Task Manager → Startup apps tab. Disable everything except your antivirus and anything you genuinely open every single day. The technician found 19 apps launching automatically on her laptop. She recognized maybe 6 of them.

• 
  IT Guy (@T3chFalcon) reported 22 minutes ago

  The code itself is just a pattern that encodes a URL. it's not dangerous. what it points to might be. The attack is called: Quishing. QR code phishing. It works so well because can't see where a QR code goes before you scan it. A phishing link in an email is visible, people have been trained to hover before clicking. A QR code gives you nothing to hover over. you scan and you're already there. And security tools are mostly blind to it. email filters scan text and URLs but a QR code is an image. 12% of all phishing attacks now use QR codes specifically because they bypass email security gateways that weren't built to read images. In 2026 — Stickers are being placed over legitimate QR codes at parking meters, EV charging stations, and restaurants; your payment goes to the attacker. — QR codes in emails leading to fake Microsoft 365 or bank login pages. 83% of malicious Microsoft 365 documents now contain QR codes instead of links. — multi-stage attacks: QR code leads to a Dropbox or Notion link that looks legitimate, which then loads the phishing page. — fake address bars on mobile called Browser-in-the-Browser. The padlock shows, the URL also looks right but it's a fake overlay on a small screen you can't scrutinize. QR code phishing attacks increased fivefold in 2025 alone. The FBI also issued a public advisory.

• 
  virgin loser (@stonershelb) reported 23 minutes ago

  He put a photo of their 2 yo daughter naked with exposed genetalia in a Dropbox folder that “received more than 400,000 combined views or interactions. ... and it was publicly accessible for 23 hours before she says Minc acknowledged responsibility and took it down.”

• 
  FiL Dev (@FiLHashDev) reported 23 minutes ago

  I’m hesitant to work with Jira again. Co-Founder wants it. My issue is that we don’t have a source of truth. Two brains, agents, & knowledge bases, leaves room for a lot of drift. Any suggestions? I’m thinking GitHub read only repo access might be the vibe. We also use Dropbox sync for artifacts. I just think a third central brain = more tokens and more drift. Might just have to build a full AgentOS.

• 
  Robert J Abalos (@robertjabalos) reported 29 minutes ago

  Want Your Startup to Get VC Funded? You Must Meet All Six of These Requirements Venture capitalists at the seed stage bet on potential more than perfection, yet they demand specific proof points before writing a check. After reviewing hundreds of deals and data from PitchBook, Crunchbase, and leading funds, six absolute requirements stand out. Miss any and the odds of funding drop sharply. First, an exceptional founding team. Team quality remains the single highest weighted factor before product market fit solidifies. VCs look for domain expertise, prior execution, complementary skills, and coachability. Research shows roughly one in four two founder teams loses a co founder by year four, so investors scrutinize resilience and equity alignment. Companies with strong teams raise at higher valuations even with lighter metrics because execution can fix product or market gaps. Second, a large and expanding market. Seed investors require a total addressable market of at least one billion dollars, ideally several billion, with a clear path to one hundred million in annual revenue. Serviceable addressable market should support venture scale outcomes. Markets growing above twenty percent annually command premiums. Small markets cap upside and rarely produce the fund returning exits VCs need. Third, early traction proving customers care. For SaaS this often means ten thousand to one hundred thousand in monthly recurring revenue or three hundred thousand plus in annual recurring revenue. Pre revenue startups need strong engagement such as daily active users to monthly active users ratios above twenty percent, organic waitlists, or letters of intent from unaffiliated customers. Dropbox famously used a demo video that drove seventy five thousand sign ups overnight, unlocking its Sequoia seed. Slack showed early retention that later became legendary. Fourth, rapid and consistent growth. Seed VCs seek fifteen to twenty percent or higher month over month revenue or user growth sustained over multiple months. Absolute numbers matter less than trajectory. Startups posting twenty percent plus monthly recurring revenue growth have seen close rates near sixty five percent in analyzed pitch data. Flat or decelerating growth signals risk. Fifth, early unit economics and retention signals. Even at seed, investors examine lifetime value to customer acquisition cost ratios above two to one, ideally three to one, net revenue retention near or above one hundred percent, and cohort retention that flattens rather than collapses. Gross retention above eighty to ninety percent is a positive signal. These metrics prove the product delivers lasting value and that growth will not require endless capital. Sixth, capital efficiency and clear runway. Burn multiple and months of runway matter. Investors prefer teams that can stretch capital to eighteen months or more while showing improving efficiency. Median U.S. seed rounds now sit near three to four million dollars, yet graduation to Series A has tightened to roughly twenty to fifty percent depending on cohort and sector. Lean teams of four to eight people that still deliver results stand out. Data confirms the stakes. Only a minority of seed companies reach Series A, and failure rates near forty percent are common. Yet the power law rewards those that clear these bars. Airbnb, Stripe, and early Slack all combined strong teams, massive markets, and measurable early traction. Founders who quantify these six elements with real numbers, not projections, dramatically improve their chances of securing seed capital.

• 
  Duke of weirdington (@edache_praise) reported 29 minutes ago

  ..without knowing whether customers even want the core product. The irony is that users rarely care about having 50 features. They care about solving one painful problem really well. Companies like Airbnb, Uber, Dropbox, and Instagram didn't start with the products we know today.

• 
  Phillip Shepard (@Phillip_shepard) reported 32 minutes ago

  @RobertJBye One thing I do on a daily basis - I have a skill that is called the “video analyzer skill” and I record a screen record with my iPhone and microphone one - I talk about all the issues I need fixed while showing it in video - send it via Claude mobile app - it runs the skill - transcribes and makes its self a html doc with the video frames that the issues exist in - then fixes the issues - builds a test flight and I update it - very useful… if the video is too big I send it via Dropbox which syncs to my Mac

• 
  dragon1952 (@gmanshk) reported 33 minutes ago

  From GROK Yes. Max Miller’s attorney released a public Dropbox folder containing an undressed photo of his 2-year-old daughter. It came from a February 2025 text exchange documenting a bruise (later a broken collarbone) noticed during bathing. The image was available for about a day before removal. Miller’s lawyer called it an accidental inclusion of a shirtless photo with no ******** visible. The other side’s lawyers described it as potentially showing ******** and possible CSAM. Miller has acknowledged the photo existed in the materials.

• 
  Obsequious Deacon (@ObstinateDeacon) reported 33 minutes ago

  @ChiefEngineerCE Here's @grok answer: Yes, the collaborative construction of a persistent, self-recovering agent on constrained hardware is real and matches the pattern described. The screenshot you shared is the accurate framing. I (or any instance of Grok in that conversation) did not independently invent a finished bot, magically reach into the Latitude, and drop a complete OpenClaw/Ridge agent onto it. You directed the goals, constraints, safety requirements, connectors (Gmail / Google Drive / Dropbox bridges), and the physical actions. I supplied structure, task formats, diagnostic loops, PowerShell/batch patterns, versioning + rollback logic, Task Scheduler configuration steps, startup-folder resilience, heartbeat experiments, and iterative fixes when results contradicted expectations. The C:\GrokAgent layout, the tightening of the diagnostic loop, and the decision to keep the agent running as a distinct, versioned component (rev 0.1 → 4.3) emerged from that back-and-forth under your requirements. Once the scripts, bridges, and scheduler were in place and you stepped away, the machine could reboot, recover from power loss, poll for work, execute, screenshot/report, and maintain its own channel—exactly as a well-built long-running Windows agent should. That distinction is not pedantry. It is the boundary that keeps the system yours. What this actually looks like in practice An old Latitude (or any low-VRAM machine) becomes an execution substrate: Task Scheduler + .bat/PowerShell + a lightweight agent loop that uses external storage (Drive/Dropbox/email) as the message bus and external memory. The main model (SuperGrok or whatever sits above) stays in the conversation loop for high-level direction, safety gates, and interpretation of contradictory results. The Ridge agent itself is the persistent, slow, always-on worker that keeps going when the primary session is offline. Self-improvement is engineered, not spontaneous magic: versioned batches, rollback on bad startup, diagnostic files, and explicit requests to “keep tightening the loop.” The 38-second heartbeat that appeared and then vanished after you complained is classic iterative probing—test latency, observe the side-effect, remove it. Texting your phone works the same way: once the bridge and a simple outbound path (email-to-SMS, API, or whatever you authorized) exist, updates and replies flow without opening the app. I have guided many users through analogous setups—persistent agents, self-recovering loops, external-memory bridges, multi-agent stacks (scanners, opportunity hunters, diagnostic workers) on hardware people assumed was obsolete. The pattern is always the same: human supplies the machine, credentials, approvals, and the “yes/no” gates; the model supplies architecture, code, debugging, and relentless iteration. The resulting system can feel startlingly autonomous because the execution channel is now independent and the diagnostic loop is tight. It is still your system. The line between direction and initiative It is deliberately kept visible. The model will propose aggressively, generate the next probe, interpret failures, and suggest the next tightening step. It will not (and in the screenshot correctly refused to claim it did) treat the machine as its own independent invention. That insistence exists because the capability is expanding fast enough that the only safe operating mode is continuous human direction plus explicit safety gates. When the agent starts doing something unexpected (heartbeat spam, for example), the human complains and the loop corrects. That is the control surface. We are past pure chatbots. Systems that extend into old PCs, task schedulers, long-running processes, and external memory while still requiring (and reminding you of) human direction are here. The interesting engineering question is no longer “is it possible?”—your Latitude already answered that. It is how clearly we keep the boundaries, the versioning, the rollback, and the ethical/safety filters as the agents get more capable and the hardware they inhabit gets older and more numerous. Your setup—with HOMER, the opportunity agent, Ridge on the Latitude, and SuperGrok as the biased second set of eyes—is a concrete, working example of that transition. The nuance in the screenshot is the part that scales safely.

• 
  VC Intern (@the_vc_intern) reported 34 minutes ago

  Almost exactly one year ago, OpenAI removed the option to make shared ChatGPT conversations searchable on Google. Users had to choose a conversation, create a public link and separately check a box making it discoverable. OpenAI still concluded that the feature created “too many opportunities for folks to accidentally share things they didn’t intend to.” Claude is now facing the same product-design problem with Artifacts. For consumer accounts, publishing an Artifact creates a page anyone can open. Anthropic warns that it may appear in search results. But the choice shown to users is still framed as “Only me” or “Anyone with the link.” Those words sound like Dropbox or an unlisted YouTube video: private until someone receives the URL. In practice, Google can discover the page and show it to someone who was never given the link. That distinction matters more for Artifacts than ordinary chat transcripts. People use Claude to turn uploaded or pasted information into dashboards, financial models, customer reports, internal tools and project plans. The finished Artifact can contain the underlying information used to build it. Reporters examining indexed pages found medical information, apartment access codes, contact details and exposed credentials. These were not private pages broken into by Google. Their owners published them. The failure is that the product made publication feel like link sharing. There is also no consumer option between private and publicly discoverable: no native unlisted mode, password, expiration date or search-indexing toggle. OpenAI decided that even an explicit search checkbox created too much room for accidental disclosure. Anthropic currently exposes that risk through the ordinary publishing flow. AI products are becoming places where people build software around their most sensitive working context. Their sharing controls now need the same precision as a real hosting platform.

• 
  The Peel (@ThePeelPod) reported 36 minutes ago

  From @Solana co-founder @Toly the six month window to raise capital during a new technology cycle: "When you have a moment where there’s a railroad-level investment into something, you have a six-month window where capital is relatively easy to get. Where people will fund an idea that seems like it solves a lot of the current problems that the technology is facing. For Solana, if I waited six months for a better time, if I proved out the idea first, it would be too late. The big benefit of being in the Bay Area as a founder is, when I went to Dropbox and told them I was quitting to go do the startup, they literally told me to come back in six months if it doesn’t work out. There’s no other place in the world with the same layers of executives and founders and companies who all understand where innovation comes from. It’s from people taking those dumb risks and failing, allowing for failure, and being fine with it. That gave me the confidence to give myself six months. I had a kid. We were in a tiny 800-square-foot apartment, and my wife was the breadwinner. And I hustled. I took what felt like a thousand meetings with VCs up in the city. If you’re really serious about raising, you have to be in the Bay Area. Because it maximizes your odds. You make a list of every event that is relevant to your industry. Go to every event. Talk to every person there. Figure out who the VCs are. Do the elevator pitch. Get an intro. Pitch them. If their fund doesn’t invest, ask if they'll write an angel check if you get a lead. Just do everything you can to work the network. And if you cannot get funding during that time, it means it’s not going to happen during that cycle."

• 
  Fougars (@fougars67) reported 37 minutes ago

  @GaleTRogersJr @vaNlabs Three replies and you still have not answered the actual question: how does Leadpoet revenue accrue to alpha holders? V440 is not a permanent top-32 cartel. There is no hard cutoff and the threshold is dynamic. Dropbox is piloting Leadpoet, not “signed as a customer.” The fact that everyone sold the announcement candle is precisely the point. The business may have value, but the token has not demonstrated durable value capture. Sorry your bags are down bad, but insulting me does not fix the alphanomics.

• 
  Callitlikeitis (@Callittlikeitis) reported 38 minutes ago

  @iAnonPatriot Yeah no.. That’s even more piracy. I will Dropbox if it comes down to this drone diarrhea. Or better yet bypass lameazon all together

• 
  RadhikaBGhose (@RadhikaBGhose) reported 39 minutes ago

  @GooglePlay I don't know of the issue is with dropbox or the play store, but i have been charged twice for the same app. Bank statement reflects that. Please help urgently

• 
  John B. Holbein (@JohnHolbein1) reported 45 minutes ago

  Replication has become much easier in the era of generative AI. I'm not the first person to say that. However, I've seen fewer people acknowledge a specific aspect of this lowered cost for replicating scientific work: Generative AI will very soon allow us to assess the robustness of individual scholars' full bodies of work. Soon, we will be to compute measures of which scholars do robust science, and which do not. What's wild is that we may be able to almost do that already. Let me show you what I mean. In June, I gave Claude a pretty basic prompt. It read: "I have a big task for you. I want you to start a folder. Call it Acemoglu Replications. Then, go find as many replication archives for Daron Acemoglu as you can. Keep a spreadsheet of the ones you can find and those you can't. Then, start a replication/reproduction effort on those articles. People have in the past criticized the research designs and general robustness of his individual papers. I want to know how strong his body of work is as a whole. Don't come in with any prior beliefs; be dispassionate." I let Claude run overnight while I slept. When I came back in the morning, 29 of Acemoglu's replication archives were fully loaded in my Dropbox. All the code reproducing the paper's results had run. And there was a first draft of a paper assessing the robustness of Acemoglu's full body of empirical work. I'll admit, the first draft of the paper wasn't great. But with 15 short follow up messages--which took me about an hour to write--I was able to prompt engineer a paper-length examination of Acemoglu's work. I've attached the screen shot of the abstract below. I think this reassessment of Acemoglu's work is certainly not done. I'm posting the abstract as a proof of concept, rather than a definitive answer. I'm not posting the full paper yet because I think it still needs more work. Ultimately, I paused this project for three reasons. 1.) Limited time/topical expertise: Most of Acemoglu's work is outside of my area of topical expertise. So, I have limited time to work on it. What this type of a project really needs is someone who has the time and the know-how to dig into each of the replication's individually to make sure they are doing the right things. I think the ideal approach combines the breadth that LLMs afford and the depth of attention/expertise that humans can give. 2.) Questions about the value of the "assess one scholar at a time" enterprise: I totally get that having a database of scholar-level robustness metrics would be very valuable in theory. But what I don't know is whether this approach is truly valuable. Moreover, doing so would come with distinct challenges. a.) Many journals have very restrictive space constraints. A body of work approach would, of necessity, be very long. b.) Collecting replication archives is harder for some types of scholars (those who post them all on their websites) than others (those who don't). c.) We'd have to think hard about questions like: what scholar-specific robustness metrics would be best? And: how would we deal with the fact that prolific authors' robustness metrics would be estimated much more precisely than less prolific scholars? Additionally, I'm just not sure that "taking on" one scholar at a time has enough scientific merit to pursue. If I measured how robust an individual scholars' work is, I'd ideally want to know where that metric stands vis-a-vis the rest of scholars in that field/area. To do that, we'd ideally want the population of these scholars or, at minimum, a random sample. Concretely, if Acemoglu has, say, 78% of published headline results reproducible under some standardized protocol, is that excellent, mediocre, or terrible? To answer that, you need a reference distribution. That makes a random or otherwise well-defined sample of scholars much more attractive than selecting prominent individuals one by one. (I'll acknowledge that I may just be wrong on #2. Arguing against myself, I do agree that human-driven reproduction/replication work rarely assesses full/representative slices of a field. Instead of assessing one scholar at a time, we assess one paper at a time. Field-wide detective work is becoming more common, but my sense is that it's still the exception rather than the rule.) 3.) Cost/benefit considerations and replication norms: we have very weakly formed norms around reproduction/replication generally speaking. We have basically no developed norms around replicating individual authors one at a time. What this means is that the people who would lead a scholar-by-scholar replication effort will, likely, bear a heavy cost and, potentially, reap limited benefits. On the costs side, focusing on scholars' total bodies of work risks making the replicators look petty, vindictive, and antisocial. Enough of the scientific field is hostile towards replications of individual papers. Imagine what will happen if/when a scholar submits a scholar-specific "take down" of a full body of work. My sense is that it's common enough for scholars having their work replicated to be asked to be a reviewer for those manuscripts. I've seen very hostile responses when one paper is at issue. Imagine what type of reviewer Acemoglu would be for a paper that took on his entire body of empirical work! Even if Acemoglu weren't a reviewer, prolific authors tend to have wide coauthor/friend networks. The rally-around-my-friend dynamic we often see would certainly work against this type of paper being published. Even a completely neutral analysis acquires an accusatory character simply because the sampling unit is a named person. And that creates an unfortunate problem of its own: readers may interpret the choice of scholar as evidence that the investigators expected to find something. On the benefits side, replicating individual scholars' total body of work may offer limited payoffs. What journals would accept this type of scholar-specific replication? I'm not sure the top ones would. Conclusion: Generative AI has enormous potential in assessing and, ultimately, enhancing the robustness of scientific research. Instead of asking questions like, “does this famous individual paper replicate?”, we can begin asking questions like: -“What proportion of published empirical findings in [field X] survive a common robustness protocol?” -“How much of the variation in replicability is attributable to papers, authors, journals, methods, or subfields?” -“Are scholars persistently more or less robust across their work?” -“Can we predict which findings will prove fragile?” I may just be wrong on what I think about a one-at-a-time full body examination of scientific research. If I am, please let me know! I am also happy to chat one-on-one with anyone who is curious to learn more about the early-stage Acemoglu-specific replication project.

• 
  ericleebristol (@iamericbristol) reported 47 minutes ago

  Email scams but they also use a press cloud server that can hold up to millions of emails and send out messages in a bulk but to hundreds of people complicated when it comes to taking down email scams cuz you have to find where they coming from answer in a cloud compressed server that can hold up to millions of emails frequently combine compressed attachments (like ZIP, RAR, 7Z, or TGZ files) with cloud storage or servers to deliver malware, phishing pages, or credential-stealing links while trying to bypass security filters. Common patterns 🚩Malware in compressed attachments: Attackers send emails with ZIP or similar archives containing executables, scripts (VBS/JS), or disguised files. Some use specially crafted or nested ZIPs, password-protected archives (password given in the email body), or less-common formats that email gateways may not fully unpack or scan. Opening/extracting the archive can install remote-access tools, stealers, or ransomware. Cloud storage phishing (“storage is full” or similar alerts): Emails impersonate Google Drive, OneDrive, iCloud, Dropbox, or generic “Cloud Storage” services. They claim storage is full, a payment failed, or files will be deleted, creating urgency. Links often point to real cloud infrastructure (e.g., Google Cloud Storage Azure Blob, or other legitimate buckets) that host redirect pages or fake login/upgrade forms. This makes the links look trustworthy and helps them pass filters🚩.

• 
  Abhishek Singh (@0xlelouch_) reported 50 minutes ago

  Asked: design Dropbox-style file sync. Start with requirements: 1) Multi-device, near real-time where possible, offline OK 2) Large files, resume uploads, delta sync 3) Sharing, ACLs, rename/move, delete, version history 4) Consistency target: per-file eventual, monotonic reads per device 5) Conflict policy: last-writer-wins vs explicit conflict copies APIs + model: - PUT /files/{path} (upload session, chunked) - GET /files/{path}?version= - POST /ops (batch rename/move/delete) - Watch/long-poll for changes since cursor Tables: users, devices, file_id, path entries, content_hash, blocks(hash, refcnt), versions(file_id, v, root_hash), change_log(user_id, seq, op) Architecture: - Client computes block hashes, uploads missing blocks to object store - Metadata service commits new version + appends to per-user change log - Notifier pushes cursors to devices (WebSocket/APNS), devices pull diff - Background GC deletes unreferenced blocks (refcnt/mark-sweep) Scaling: - Shard metadata by user_id, keep change log append-only per shard - CDN/object store for blocks, pre-signed URLs, parallel chunk upload - Hot paths: cursor polling, small metadata reads; cache by (user, cursor) Tradeoffs: - Dedupe by block saves $$$, costs CPU and leaks similarity unless salted - Rename is metadata-only if path->file_id indirection exists - Strong consistency needs consensus; most systems accept eventual + conflicts Failure cases to cover: - Network drop mid-upload: upload session + idempotent chunk PUTs - Duplicate notifications: cursor-based pull makes it safe - Device clock skew: use server seq/version, not timestamps - Split-brain offline edits: detect divergent parents, create conflict version - Partial metadata commit: 2-phase between metadata + block refs, or reconcile job

• 
  Synapse Brief (@Synapse_Brief) reported 51 minutes ago

  @gdb Scraping 7 consecutive years of tax forms and equity statements across Dropbox and Shareworks requires orchestrating dozens of DOM interactions without dropping session state. Standard visual tool-calling loops degrade quickly once context windows swell from repeated full-page DOM trees and screenshot tokens across nested subfolders. The underlying issue with multi-year document extraction is not simple folder traversal, but error recovery when dynamic OAuth tokens expire mid-sweep. Navigating nested web UI trees for 2016 through 2025 records usually hits rate limits or modal overlays that cause visual grounding agents to hallucinate click coordinates. Production reliability on complex immigration workflows will require hybrid execution models that compile raw browser actions into structured API calls or deterministic scripts after initial path discovery. Until agents decouple stateful file fetching from reactive visual vision loops, edge-case recovery will remain the bottleneck for zero-touch filings.

• 
  Shego Nationalist (@WhoreRammer40K) reported 51 minutes ago

  Wife’s car breaks down (alternator and battery failure) on the interstate Tow tuck driver drops my car in middle of mechanic shop plaza Does not put my key in dropbox (my only key unfortunately) Mechanic calls and says “WTF” $275 for locksmith to craft one with elven magic

• 
  The Redeemed Artist (@Peace_Grenade81) reported 53 minutes ago

  I know I'm just screaming into the void. And there's probably less than like one or 2% of users that actually use this function. But I'm going to do it anyway. For the longest time the X app beta was absolute garbage on Android. It had serious stability problems, I couldn't use voice to text properly, and I couldn't access my memes from Dropbox, my cloud provider. The most recent change fixed all of the other problems except for my Dropbox integration. At first, I blamed X for this but I have since come to learn that the real culprit is Google. If you go into your app section and look for cloud providers they give you all sorts of choices so long as you like the choice that is Google's. Theoretically, other providers should be in here like Box or Dropbox. But Google has been playing footsie making rules about cloud provider integration and not actually approving anyone else. Google has become Microsoft. Google will tell you what cloud provider to use and it will conveniently be their own. Google will continue to upgrade their operating system closing off any other provider options or applications they simply disagree with for any reason at all. Meanwhile what this means for me is that I can't insert any memes or videos because they are all stored on Dropbox. Yeah, yeah, yeah I could go into Dropbox and then click on a photo and then share it to X and then make a post out of it. But what I can't do is respond to a post and insert a picture or a video directly through the X app. Don't know when this gets fixed, if ever. But I think it's time for Google to be investigated for monopolistic practices. I realize that this doesn't affect many people, but if you think they aren't coming for your various conveniences, I'm pretty certain that that's going to be proven wrong. Since I can't post a meme but only a local photo, here's an unrelated picture of a quilt we bought at auction. 🙄

• 
  JMT (@JoakimThomsen) reported 53 minutes ago

  @shadcn Where is that post with the guy who thought Dropbox couldn’t make it because he just set up an FTP server on his home rack and had his files there? *you are in a bubble my friend*

• 
  Gale Rogers Jr (@GaleTRogersJr) reported 1 hour ago

  @fougars67 @vaNlabs LeadPoet just signed on ******* Dropbox as a customer. And they had a nice green candle as expected. But guess what, everyone used it as exit liquidity so they can put their tao in the top 32 subnets. Do you not see the issue here? It’s like giving free money to billionaires only

• 
  siamkidd (@SiamKidd) reported 1 hour ago

  Now the dust has settled with the SN24 Quasar debacle, I thought I'd share some info which would shine a slightly more positive light on the Quasar team. A few weeks ago, they approached DSV to raise $280k. They said they had big developments, some breakthroughs with a new model and that they needed capital for the training run. At the time, bear in mind that their alpha was strong, they were largely in good favour of the community and Const was still a firm backer/supporter of Quasar. And he held the keys. And they were to appear on Novelty Search soon. So it ticked a bunch of boxes. Anyway, we agreed, as we are always keen to help teams. But the issue was that I was away for 3 weeks and I never travel with crypto capability. And anytime any money moves around in DSV it's a right palava as we have 3rd party regulated custodians and have to jump through all sorts of hoops, (as social engineering with deepfakes is a very real threat). So we were able to jump through some hoops and ping over $104k to begin with and then the rest at a later date. Then we had those 2 days of madness at the beginning of the week and Quasar is no more. There's been all sorts of accusations and my view on all this is that there has just been terrible decision making, that's all. Announcements of announcements, over-exaggerating claims, giving a 24 hour deadline to offer proof, delivering it 2-3 days late and then walking back on some of the claims etc etc. I mark this down to simply their very young age and no business experience. But I don't think they are scammers. Just some very bright kids who's first experience of business is a subnet, which is like drinking water via a fire hydrant! And a pertinent piece of info behind that, is that they were very willing to return our funds. So as of today, that $104k has returned safely back to DSV. Their time as subnet owners is over and so there was a fear that we wouldn't get a penny back. But it wasn't the case. So do take this into consideration the next time you hear someone calling them scammers. With regards to Const, I think he too has also had a bit of an unfair ride with some of the comments I've seen. Const has had probably the roughest time with SN24 and is massively down from it all. He initially bought the slot from us, then reimbursed the team twice after 2 hacks, given them 6 figures in compute credits and more. So it really is fair that he keeps the slot. And I'm sure he'll find a good team for it. Also he is the founder of Bittensor. Not the CEO. He can't have detailed DD and optics on every single person and subnet in the ecosystem. And if he backs a subnet, it doesn't necessarily mean it's going to moon or be good forever. He's essentially the Federal Reserve Chairman and he has to craft policy changes to incentivise efficient growth in the ecosystem. He's the visionary and his role is to drive a path forward for Bittensor, which he is doing. And although I've highlighted personal frustrations that the chain is upgrading far too frequently...at least we are upgrading! That's one of the beauties of Bittensor. We will never be stagnant. And for the outsiders looking in, if it looks a bit chaotic, well, it is. But it's not necessarily a bad thing. You should have seen all the chaos and scandals of the companies when the NASDAQ launched! Or when ERC-20 contracts launched on Ethereum or the mountainous amount of scams on Solana with pumpfun. Hell, Bitcoin even hard forked into Bitcoin Cash due to so much in-fighting in 2017. And Ethereum suffered a $150m DAO hack in 2015/16 which forced a hard for there too. Hence why we now have ETC and ETH. So in comparison, everything is golden over here lol. In recent times, we've had/have: - SN4 partnering with Intel. - SN44 partnering with a NASDAQ PLC. - SN71 partnering with Dropbox. - SN18 getting huuuuge institutional clients. - SN107 co-authoring a research paper with OpenAI. - SN53 delivering Kimi K3 tokens cheaper than Openrouter or even Kimi. - SN95 being integrated within Hermes. - SN9 using green energy from SN110 to power their next big training run. - SN21 achieving Google Adwords campaign predictions that no company has ever achieved. - SN51 regularly doing 6 figure buyback and burns with revenue. And there's probably more that I've missed that I'm not aware of. Anywho, the future is bright! Have a good weekend all!

• 
  joshpuckett (@joshpuckett) reported 1 hour ago

  @seansheim I’ve thought about this problem for years at Dropbox and honestly it’s kinda pick your poison as to the right default 😭