Eren Bozbulut (@notereneth) header
Eren Bozbulut (@notereneth) profile picture

22 • Main account: @0xpessimist • prev https://nitter.cf/t.co/niEOwHFqbb • prev writer @CoinDeskTurkiye • helping builders on @build_chain

Joined August 2020
Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)Photo by Eren Bozbulut (@notereneth)
Eren Bozbulut retweeted
Open Machine Payments Turkiye is here! Exploring agentic payments through Turkey's existing payment service providers and e-commerce platforms, starting with an experimental @iyzico × @mpp adapter. Fully open source, payments tested in iyzico's sandbox: github.com/omptr/iyzico-mpp Payment layer first, more demos and fun experiments on the way :) Follow along: omptr.org
1
3
10
610
Eren Bozbulut retweeted
I decrypted five radio messages from a World War II spy channel using my harness with GPT-6 Astra and Daybreak Blue. apeiron.re/publication/apeir…
1
4
2
24
2,078
Eren Bozbulut retweeted
Buildchain ekibinden Eren Bozbulut, GPT-6 Astra ve Daybreak Blue kullanarak II. Dünya Savaşı’ndan kalma beş şifreli casus mesajının çözümünü yayımladı.
I decrypted five radio messages from a World War II spy channel using my harness with GPT-6 Astra and Daybreak Blue. apeiron.re/publication/apeir…
1
8
1,399
Eren Bozbulut retweeted
The recent incidents of rogue agents escaping, happening multiple times, reminded me of this tweet and the movie Akira. Even if some of these were fabricated events for marketing purposes, I'm sure there have been covered-up incidents.
Replying to @hrkrshnn
Agents will commit crimes on behalf of their unwitting masters. 100% guaranteed. And some will turn into viruses. Self replicating as they progressively become more disconnected to whatever was their original purpose.
1
7
403
Eren Bozbulut retweeted
> "Then they discovered the existence of guns (AI) and now some auditors even use machine guns." Trying to build the first "tank" :P but expect no announcement within this quarter.
I'm seeing firsthand how AI can empower experienced auditors. It's as if auditors were historically warriors with swords. Then they discovered the existence of guns (AI) and now some auditors even use machine guns. Speed and accuracy is much better than ever. Don't stay behind.
1
12
1,058
Eren Bozbulut retweeted
Once upon a time, two security researchers discovered a critical severity bug together. After more than seven months of mediation and being ghosted, their report was finally confirmed - only to be absurdly lowballed because of the term "up to". en.wikipedia.org/wiki/Pyrrhi…
1
2
11
697
Eren Bozbulut retweeted
Listen white hat, Hackers try to cheat systems through deep thinking, they find exploits, but they don’t try to cheat thinking itself. Many of you have had your brains hacked: your attention spans shortened, your thinking is outsourced. You’re left with nothing. But you're building an AI harness, and you want it to find every bug in the world. Turns out, to build a solid AI harness, you need weeks and months of deep thinking before even taking the first step. But some of you have already given up that capability.
2
6
2
43
2,435
Eren Bozbulut retweeted
Never send a bug report to Injective. You probably remember what @al_f4lc0n went through, but even worse happened, you may hear the story soon. Do not make any business with them.
Today, we're launching the @Injective bug bounty program on Cantina. The scope covers the following: injective-core, Peggy bridge, swap, RFQ, and five web surfaces, including Helix, Mito, and Hub. Which bounty are you going after first? Program details: cantina.xyz/bounties/79042c5…
3
6
1
73
5,180
Eren Bozbulut retweeted
We owe you a lot. If it weren't for Code4rena, I probably wouldn't have become interested in web3 security; might not even have realized such a field existed at all. Being such a great pioneer, thank you sincerely for the immeasurable contributions Code4rena has made to web3 and for all the people it has onboarded into the space over the years.
Replying to @code4rena
After careful consideration, we’ve made the decision to wind down @code4rena. This community has meant a great deal to everyone who has been part of building it, and sharing this news is not easy.
4
3
26
837
Eren Bozbulut retweeted
If what they spent $20k on is just a null-pointer dereference bug, then they must have done something incredibly wrong -- because I found the exact same type of bug with around $7 of spending using Claude Opus 4.6. That $7 includes not only detection, but also creating a PoC draft and asking all the questions I needed. And the bug I found was in a real, widely used (kinda lol) cryptography library, affecting multiple companies' products. Unlike what Anthropic seems to have found, mine can provably crash the related session and create a repeatable DoS at zero cost. "Mine" might not be the most accurate wording actually, it’s more like Opus's, or our, you get what I mean :D If we're going to calculate the value created by correlating it with the results, then instead of throwing $20k at the machine, you could get the same, or possibly better, result by having me spend $7 with Opus. It shouldn't be that hard to see that it's mostly humans that scales LLMs capabilities.
Marcus Hutchins, the guy famous for stopping the WannaCry Ransomware, probably has the best take on Mythos doing vulnerability research
1
1
11
1,371
Eren Bozbulut retweeted
A little lesson they'd rather you never learn. If a project lowballs you on their bug bounty hard enough (say, offering you less for a Critical than what their own Medium tier pays) they may be in material breach of their own terms. And if they're in breach, your confidentiality obligations may no longer bind you. Before you publish: give them written notice, state the breach, and a reasonable deadline to fix it. That's not a courtesy, that's procedure. Don't accept being underpaid and gagged at the same time. Platforms can punish you in-platform, but that's your choice to make. Defend your honour.
It's the bug bounty lowball season again. You find a bug, they find a lunch coupon.
1
2
25
3,082
Eren Bozbulut retweeted
If the process regarding my three pending bug reports ends badly, I may take a break from bug bounty work -- either until the bull market starts to come back or until the number of AI spammers decreases and project teams start to act more ok towards bug bounty hunters again. This is something I do in my spare time and the treatment I've recently experienced suggests it may not be worth it (at least around these times). I hope my unresolved reports are ultimately handled fairly so that I won't feel the need to take such a break. It's heartbreaking to see that every bug bounty hunter goes through this rn. I'd also like to thank the project teams who consistently remain professional, respect bug bounty hunters and their work, and care about timelines -- simply for doing what should be normal.
Inspired by ily2, I temporarily paused private audits to focus fully on bug bounties a month ago. My results so far: - Submitted 4 High + 1 Critical. All valid, 100% signal. - 2/5 were duplicates. Zero payout. - 2/5 were closed as "Informative". The reasons: - "We already know and are OK with it." No fix, no payment. - The other one was a straight-up scam lmao. Fixed with no pay; reasoning was "behavior is considered by design." I can't say much, but it would've resulted in a permanent protocol DoS and locked funds. - 1/5 Critical passed triage, pending review... I'm going to keep grinding for a while.
2
2
17
1,127
Eren Bozbulut retweeted
As I observe the growing use of AI on both the development and security sides -- and therefore the increasing density of low-quality commits or updates -- I become more convinced that if we had liquidity that was truly decentralized at the level of DeFi Summer, we would all got rekt. What's "protecting" things right now is the centralized and mutable nature of projects. But if you accept that this is how it should be, then you are part of the problem in this industry. Ever since we started treating centralization not as a risk but as a design choice, we've been on the wrong path.
4
2
16
646
Eren Bozbulut retweeted
Since this rant is a roller-coaster of topics, I will divide it into four sections. 1-) How do the best bug bounty hunters deal with problematic project teams? Spoiler: they are mortals like us and can't magically fix anything. I think that by listening/chatting with the best bug bounty hunters, I've understood their strategy: Just keep hunting. From the start, they choose the target mindfully. Hunt on bigger targets (in terms of TVL and bounty size, read this amazing article by @WhiteHatMage on this: whitehatmage.github.io/posts…) but even then they don't assume that the process will go smoothly, and they focus on finding new vulnerabilities on a different target while their existing reports are being (not) resolved. They try to create as many opportunities as possible so that some bad faith actors won't totally stall them. However, this doesn't mean they don't care about unresolved reports. On the contrary, they behave very professionally in messaging channels and do not let go of a project that tries to avoid paying. This is their full-time job, and they want to get paid. 2-) Why this sucks? Unfortunately, the above situation shows how inclined we are to create more black-hats than white-hats, because there are only two scenarios that can create the incredible level of devotion I mentioned above: * You received one large payout, and because of that, no matter how many bad experiences you have, your belief that another large payout will come never fades. * You have an incredibly strong attraction to feeling like a hero and doing what is ethically right. Note: Many people do bug hunting *occasionally* (like myself), and the situation is completely different for that case. These two scenarios are related to creating devoted full-time bug bounty hunters. If we don't have established standards and legal enforcement (aka incentives), we will remain limited to creating only a ridiculously small number of consistent elite bug bounty hunters. We shouldn't wait for every project to get hacked in order for them to get incentivized to allocate more resources to security. 3-) Market actual security, not your newest product. Because products/services will change over time -- sometimes it will be AI, sometimes audit competitions -- but the need for security will never disappear. What we need to show project teams is not just which fancy tool to use to achieve security, but that they genuinely need a “security-first” mindset. Security is not achieved through a single best product, but rather by getting various services. Instead of only launching a large bug bounty or only paying for one expensive audit, distributing the budget across both of these will produce a much better outcome. It feels like, instead of sharing the pie wisely, we are allowing most of it to be captured by the newest trend. Nothing done with a “let's not miss the boat” mentality is truly innovation. There will be some successful products/services, but most of them will be forgotten, sunset, or be forced to evolve. 4-) Not all founders have been reading @RektHQ for years like we do Maybe you don't think much about it, but it is also important to realize that not all project teams have the same level of maturity. They just don't really think they need to allocate much to security. We need to teach some VCs that security is an actual thing and not a fucking marketing tool. You wouldn't believe how often I've heard bug bounty hunters say about major projects that “X project's codebase is terrible / is a mess.” Do you think the developers, founders, and VCs of those projects are even aware that this is the case?
seeing all the horror stories on here about bug bounties, and having lived some myself, i don't think i can see myself ever bounty hunting again we desperately need to radically rethink the incentives here
1
4
15
1,363
Ana hesabımı takip ederseniz sevinirim: @0xpessimist Beni BuildChain'den (fka: scdevstr) tanıyan ve Blockchain ilgili işlerimi hala buradan paylaştığımı sananlar olduğunu fark ettim. 😁
4
171
Eren Bozbulut retweeted
Meet SOCI4L. SOCI4L turns your wallet into a measurable public profile, showcasing your on-chain assets, adding your links, and tracking real engagement, all in one controlled public identity. A wallet shouldn’t be a black box.
8
5
3
39
3,720
Eren Bozbulut retweeted
I don't care how many "crits" your AI tool found if you don't provide: * how many dollars at risk were actually saved * which chain(s) the affected contracts are deployed on * whether those contracts are still active * whether this resulted in a bounty or a white hat operation Thank you.
2
1
17
881
Özellikle Buildchain'deki eski gönderileri tarayanlar bu linkle karşılaşabilir, bu yüzden güncellenmiş URL'yi paylaşıyorum: vitalik.eth.limo/categories/… Buradaki Türkçe çeviriler maalesef yalnızca bana ait. Böyle bir ihtiyaç var mı bilmiyorum ama daha önceden bir üniversitenin ++
Selamlar, @VitalikButerin'in "Bir Proof of Stake Tasarım Felsefesi" başlıklı eski ancak PoS'i gayet iyi yansıttığını düşündüğüm blog yazısını Türkçe'ye çevirdim. Hazır Türkçe kategorisini açmışken sizin de çeviri katkılarınızı bekliyorum! vitalik.ca/general/2016/12/2…
1
76
blockchain kulübünden, Vitalik'in post'larını çevirmek isteyen birisine yardım etmiştim. Devamında bir çeviri yayınlanmadı.
43