BBGtopshot (@BBGtopshot) header
BBGtopshot (@BBGtopshot) profile picture

Blockchain and JPEG enjoyer🤝

Joined November 2021
Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)Photo by BBGtopshot (@BBGtopshot)
BBGtopshot retweeted
Formally verify your smart contracts
🧵 Formal verification on @solana Absolutely thrilled by this 1st release. @HarmonicMath, founded by @vladtenev and @tachim, just added support for Rust to their product Aristotle, with Solana programs being a FIRST-CLASS citizen. They worked closely with 15 teams from ecosystem, calibrating Aristotle on the typical properties you'd expect from an on-chain program on Solana
140
51
15
681
82,105
Bitcoin is available on Liquidium for @oisy users! Manage all assets via one wallet. It doesn't get easier!
4
13
63
931
BBGtopshot retweeted
You type the amount and OISY Smart Swap gets the best rate. OISY Swap compares the swap providers' offers in real time and picks the best one for you. All you have to do is review it and confirm it. Try it on oisy.⁠com
3
26
1
115
1,636
BBGtopshot retweeted
Cloud engines went live on Friday. They're AI-native by design. A web app on an engine works like a mobile app: memory, auth and hosting all come as one thing, so agents have way fewer pieces to wire up. Here's mine running a few of my apps 👇 @dfinity
4
20
90
1,449
BBGtopshot retweeted
I built my own private cloud drive on @dfinity ☁️🔒 CloudSef: folders, QR sharing, zip backup. Files and names get encrypted on my phone with vetKeys. Not even the canister can read them. My files, my keys. Idea to live in under 24 hours. What would you keep in yours?
7
21
126
2,223
BBGtopshot retweeted
As predicted, AI continues to accelerate the security challenges facing anything built on traditional tech. We've predicted this for years, and it's one of the reasons for ICP cloud engines → build using tamperproof code.
Threat actors are weaponizing ARTEX, an open-source, large language model (LLM)-driven agentic platform to automate multiple phases of cyber intrusions at machine speed. Since at least August 2026, a variety of malicious cyber actors have used ARTEX to conduct campaigns targeting organizations globally across multiple sectors, including financial services, government, and technology. ARTEX significantly reduces the time required to chain attack techniques and enables threat actors with minimal technical expertise to conduct complex, multi-target operations simultaneously. Today, the FBI released a FLASH highlighting the tactics, techniques, and procedures and indicators of compromise associated with the ARTEX framework. Organizations should follow the mitigation steps outlined in the FLASH to improve their security posture in response to ARTEX and similar agentic SI frameworks: ic3.gov/CSA/2026/261113.pdf
47
205
9
798
19,514
🧑‍💻 Here are crypto’s top 10 AI & Big Data projects by development. Directional indicators represent each project’s ranking positioning since last update: ➡️ 1) @dfinity $ICP 🥇 📈 2) @origin_trail $TRAC 🥈 📉 3) @chainlink $LINK 📉 4) @nearprotocol $NEAR 📈 5) @filecoin $FIL 📈 6) @swarms_corp $SWARMS 📈 7) @oasisprotocol $OASIS 📈 8) @phalanetwork $PHA 📈 9) @vana $VANA 📉 10) @livepeer ethereum:0x58b6a8a3302369daec383334672404ee733ab239 📖 Read about the @SantimentData methodology for filtering notable github activity data from project repositories, and why it is so useful for crypto trading: medium.com/santiment/a-diffe… 🔖 Bookmark our AI watchlist here, and see what others in crypto can’t: app.santiment.net/screener/a…
11
42
9
171
20,456
BBGtopshot retweeted
EXCLUSIVE: Dominic Williams on Cloud Engines, AI Agents & ICP’s Next Chapter We didn’t let AOL own the internet. Why did we let three companies own the cloud? In this interview, I sit down with DFINITY founder Dominic Williams (@dominic_w) to discuss Cloud Engines and his bet that AI agents will change how software is built, and which infrastructure wins. After 46 years of writing code, his verdict on his own role is striking: “I’m a super expert software engineer. I’m obsolete.” We get into sovereign cloud infrastructure, OpenSaaS, Pakistan’s early work with ICP, and the proposed model connecting Cloud Engine payments to ICP burning. As Cloud Engines enter early access, this is the thinking behind the rollout. Watch the full 42-minute conversation. @dfinity Recorded in Dubai ahead of the rollout. Chapter timestamps below.
40
190
16
593
129,347
Next week, I'm releasing an episode many of you have been waiting for. I visited the headquarters of @dfinity in Zürich, and recorded an 80-minute podcast with the founder of $ICP, @dominic_w. A thoughtful, interesting conversation where we talked about the challenges humanity currently faces as AI rises. I've learned a lot from the episode, and I'm really keen to hear what you all think. Releasing on Thursday!
51
114
7
562
48,172
BBGtopshot retweeted
十年前,我开始使用 1Password。但我当时有顾虑:掌握后台权限的人,是否仍有机会绕过授权,拿到我的密钥? 于是,我用 Go 写了一个团队版密码管理器。功能太简陋,放弃了。其他开源免费的替代品,我也没敢用。一直用 1Password,太权威。 三年前,我接触到 ICP,一种去中心化的计算基础设施。我觉得,终于有机会从架构上移除 admin 特权了。于是创建 ICPanda DAO,开始探索,后来还 @ 过 1Password。 两年前,我在 ICP 上开发了 dMsg.net,一个后端完全运行在链上的 E2EE 消息系统。但它依然太简陋,也很难用。 今年,Fable 5 问世后,我彻底认可了 AI 的编程能力。我用 AI 重新实现了 dMsg,把它 Vibe 成了自己一直想要的样子:端到端加密、保险库多端同步、安全文件分享、AI 访问权限管理…… 文章通过 dMsg 的加密文件链接分享出来,里面写了我的思考,欢迎读一读,亲自体验一下。
Who gets access to your private data—and who can change the rules? A contract you share with a client. A password you need on another device. Project files an AI assistant helps you work with. Each task involves an access decision. Behind it sits another question: who has the power to change the software enforcing that decision? We’re building dMsg as a private workspace for passwords, notes and files, with encrypted vault sync across authorized devices, selected file sharing and explicit permissions for connected AI agents. Its trust model has three parts: • End-to-end encryption protects private content. • ICP canisters enforce critical account and device authorization. • ICPanda DAO’s SNS governs upgrades to those canisters. The practical difference: an operator cannot use access to dMsg’s off-chain cloud service alone to rewrite those authorization rules or enroll a device with vault decryption access. Adding a device record is not the same as giving that device the keys. Valid authorization and the appropriate encrypted key material still matter. For AI, permissions should fit the task. Authorized operations can run within their granted scope; requests for your formal signature still need approval. Content you allow an external AI client to read also enters its own data-handling environment. Governance can approve upgrades. Client software, recovery and the ICP network remain parts of the trust model. We explain those boundaries alongside the design. Our new article, “Private Data in the AI Era,” explores these choices, with links to the protocols, source code and governance records. We’re sharing it through dMsg, so reading it is also a chance to experience receiving a shared file. Read the full article in your browser. The link below is a dMsg encrypted file-sharing link. dmsg.net/s/4017ad4fcb469a071…
4
8
1
38
1,263
BBGtopshot retweeted
I have a major result to share. I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China. This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model. Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves. I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems. I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used. It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20. These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs. I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at unemploymentlabs.com/checkpo…. There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
1,819
6,427
2,198
41,229
6,726,896
BBGtopshot retweeted
I guess that leaves only the Internet Computer in the "world computer" game then, building on hardcore technical work since 2014 ;) Cloud engines this week ✌🏼..
Expanding on why I believe "ETH as store of value (SoV)" makes much more sense to focus on than "ETH as fee-burning token". Three primary reasons: 1. ETH as fee-burning token never was as strong as it seemed 2. Optimizing for ETH as fee-burning token would demand Ethereum and ETH function in ways that are very hard to make compatible with decentralization and capture resistance. 3. Meanwhile, optimizing for ETH as SoV is aligned with Ethereum focusing on its unique strengths (decentralization and CROPS) EIP-1559 (including its burn) is fine and good, we should keep them. But we should not feel sad if fee burn is < 10% of breakeven indefinitely; the value of ETH comes from elsewhere. The core idea: the big thing that people appreciate about Ethereum is decentralization and CROPS. The main pragmatic benefit that these things translate into is stability. SoV is the category where that is not an unneeded extra we merely put effort into for ideological reasons, but rather a pragmatic strength from the perspective of people putting in their assets. And in the era of AI disruption, stability will be highly valued.
88
306
44
1,125
49,359
BBGtopshot retweeted
🚨 DFINITY Foundation has begun sending out "cloud engine" coupons. We're going slowly, but will go faster in coming weeks, as we process a large waitlist. The timing of the massive Microsoft outage is coincidental, but current happenings are part of the why. 🧵
145
520
94
1,609
85,769
BBGtopshot retweeted
dMsg is now on the App Store 🐼 A private workspace for your passwords, files and messages—now on iPhone and iPad. Built by ICPandaDAO, powered by Internet Computer. Try it and tell us what you think: apps.apple.com/app/id6827050…
14
33
6
146
3,633
BBGtopshot retweeted
Appic is now Swapic 🔁 Swap any token across 14 chains in one step. Sign once the rest we take care of. Gasless everywhere. Settled onchain by ICP canisters. Cross chains, not fingers. swapic.xyz
8
5
1
17
654
BBGtopshot retweeted
This is getting ridiculous. If any centralized party needs tamperproof custody technology that cannot be hacked (only requires those in approval chains to do their job) then DFINITY can help.
LATEST: French bank Crédit Mutuel Arkéa confirms it was hacked for $133M+ in crypto, with @zachxbt tracing $108.8M in $USDC withdrawn from @Krakenfx, @cryptocom and @OKX accounts over 21 hours, then converted to $ETH to avoid stablecoin freezes.
118
267
27
1,009
36,341
BBGtopshot retweeted
We will all be learning in coming weeks & months, that most human-made software: ▫️ Was really slow ▫️ Had many security vulnerabilities ▫️ Was under-specified and incorrect AI will rewrite it all. Many times. It will make it fast, cost-efficient, and secure. It will think about failure modes we couldn't wrap our heads around. It will prove, formally verify, load test, harden. It will turn memory-unsafe foundations into robust, optimized ones. If you ask, it'll thoroughly document it, teach you about it, make diagrams, animations, and slide decks. In many ways, we are witnessing the true birth of software engineering.
548
743
263
7,683
354,131
BBGtopshot retweeted
binance losing $6.3B reportedly started with one compromised devops laptop, which is the exact reason i'd rather trust ICP canisters than a custodian how liquidium․fi holds your collateral: 1. no custodian, canisters on ICP hold the assets 2. key shares for native BTC are split across ICP subnet nodes 3. trail of bits reviewed the canisters still smart contract risk and nothing's unhackable, but one laptop shouldn't be able to drain the vault
3
6
52
1,372
BBGtopshot retweeted
New Anthropic research: progress on provable inference for large language models. Here, we introduce FlashProve. Building on established cryptographic techniques, FlashProve allows developers to verify whether a given response came from a given AI model. It’s substantially more efficient than prior open-source methods for doing so. These proofs have a range of applications. They can, for instance, make it harder for attackers to hijack a chat, or enable safety evaluators to verify that the model they’re testing is the same one being served to customers. For our full write-up, see here: anthropic.com/research/prova… Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that end, we're open-sourcing FlashProve’s code, here: github.com/anthropics/FlashP… We’re grateful to our collaborators: Alessandro Chiesa, @GiacomoFenzi, @gwberg95, @miranda__christ, Rahul Dalal (rdalal.github.io) and Ben Mayer (@transmissions11).
655
546
504
5,293
753,089
BBGtopshot retweeted
decentralization - don't call it a comeback from today's SEC interps
25
16
8
161
16,546