
Tech consultant & anarchist technologist exploring network society theory, blockchain infrastructure, and AI systems. @BariaDAO founder
earth
Joined March 2009
- Tweets15.9K
- Following4K
- Followers2.4K
- Likes19.5K
bparlan 🦇🔊 retweeted
Tarık Pasha tutuklanarak cezaevine gönderildi. Haberi Jahrein verdi.
Jahrein: “Bir haber aldım. Sizinle paylaşmamı abisi istedi. Sayın adalet bakanıyla ilgili bir video çektiği için Tarık Pasha dün gözaltına alınmış, Ankara'ya götürülmüş, bugün savcı karşısına çıkmış ve tutuklanmış. 18'ine henüz yeni girmiş bir genç. Biliyorsunuz, benle ilgili de videoları var. Beni o komik bıyıklı adama falan benzetti. Kendisine çok kızdım ama ben o zaman da söyledim; iyi niyetli bir çocuk olduğunu düşünüyorum. Abisiyle görüştüğümüzde ilk adli kontrol olur diye düşünmüştük ama tutuklama oldu.”
bparlan 🦇🔊 retweeted
Bu ülkede yine vatanına aşık hiç bir partiye mansup olmaksızın malum süreçlere, haksızlıklara, yolsuzluklara ve bu diktatörlük rejimine genç yaşında olmasına rağmen karşı durabilen Tarık Pashamız gözaltına alındı.
İfade özgürlüğümüze set vuranlar hepiniz hesanını vereceksiniz.
The competitive moat that Anthropic & OpenAI have is completely collapsing.
AI is becoming commoditized faster than any technology in history!
Token costs are falling at a pace that dwarfs electricity, computing & even DNA sequencing.
$1.4 TRILLION valuation for Anthropic...
bparlan 🦇🔊 retweeted
BREAKING: The US Social Security Trust Fund’s deficit jumped +$39 billion YoY in Fiscal Year 2026, to a record -$221 billion.
This fund finances retirement benefits and payments to eligible survivors of deceased workers.
This comes as the fund’s costs, primarily benefit payments to retirees and survivors, soared +$83 billion YoY, to $1.5 trillion, an all-time high.
At the same time, income, primarily from contributions to the Social Security system, rose +$44 billion YoY, to a record $1.3 trillion.
Since FY2022, the fund’s deficit has surged +$188 billion, or +585%.
As a result, trust fund reserves fell -$221 billion YoY in FY2026, to $2.18 trillion, the lowest since 2007, down down -$640 billion from the 2017 peak.
The financial health of Social Security is deteriorating rapidly.
bparlan 🦇🔊 retweeted
Hey @Grok, concise no spin.
According to an ex-CIA agent, President Trump confessed 10,000 JFK files would never be released because “every single one of them points directly at” what?
Name only
bparlan 🦇🔊 retweeted
If you think your investment decisions are bad...
🇹🇷 Turkey has sold $36 billion in gold and foreign reserves this year to protect its currency, the Lira.
And the Lira chart looks like this.
bparlan 🦇🔊 retweeted
Red Team screwed us, let's give Blue Team a shot.
Blue Team screwed us, let's give Red Team a shot.
Red Team screwed us, let's give Blue Team a shot.
Blue Team screwed us, let's give Red Team a shot.
Insanity: doing the same thing over and over and expecting a different result.
bparlan 🦇🔊 retweeted
They will keep printing until fiat currency loses all value.
Democrats or Republicans, the outcome is the same.
Buy Bitcoin.
bparlan 🦇🔊 retweeted
Devletin ücretsiz yapay zeka altyapısı (inference + gpu) sağladığı platformda blockchain entegrasyonu olduğunu biliyor muydunuz?
bparlan 🦇🔊 retweeted
Compiled a list of protocols that have shut down in 2026.
82 protocols year-to-date.
I've never seen anything like this before.
bparlan 🦇🔊 retweeted
İskan'da çalışırken onlarca boş görünen dükkan ve arsayı kendi ailesi adına çeviren memurlar duyduk gördük. İnşallah hepsi araştırılır. Hatta oradan ayrıldığı halde içerideki tanıdıkları ile boş yerleri sorgulatıp üzerine almaya devam eden de biliyorum. Sonra nereden gelir bu kadar zenginlik kıçı kırık insanların? diye sorarız
📢 İSKAN KOMİTESİ BAŞKANI SAYIN AKTUNÇ
⭕️ Konut vasfı taşımayan bir yapının yasaya aykırı bir şekilde İskan Dairesi personeline kiralanması yönünde karar aldığınız; hemen ardından ise İskan Dairesi Müdürü’nün söz konusu yapının “yaşanmaz” durumda olduğunu belirterek belediyeye yıkım talebinde bulunduğu belgelerle ispatlanmıştır.
⭕️ Kamuoyunu derinden yaralayan bu çelişkili, usulsüz ve yasa dışı uygulamalar ve kararlar hakkında ne zaman resmi bir açıklama yapacaksınız?
bparlan 🦇🔊 retweeted
Sevgili @Dramatoryum Ali hocam ve @cenk_tarhan Cenk bey, Türkiye'nin en değerli teknoloji sayfası @TechnopatNet 'de oyunuma sayfa ayırmışlar, sağolsunlar. Çok gurur duydum, mutlu oldum. Yolu açık olsun inşallah. 🙏
technopat.net/2026/10/14/tur…
@pixcodream #ConvoyInc #gamedev
A Russian lab worker has died and thousands are in quarantine after a plague outbreak.
The World Health Organisation says the pneumonic form of plague is especially contagious and "can trigger severe epidemics".
@SkyNewsThomas explains everything you need to know about varieties of the disease.
🔗 trib.al/6kuLirx
bparlan 🦇🔊 retweeted
Herkes ayrı yerinden okuyor, seçim yasakları arifesinde okursak, bence Meryem hanım, seçim yasaklarını delecek haberleri yapıp hapse girecek yeni müdür bulun dedi, çünkü çok belli ki gene ve yeniden seçim yasaklarını BRT'ye deldirtecekler, bence bu istifa bunun işareti
bparlan 🦇🔊 retweeted
Rusya’da %40 civarında kalan katılım oranı da, %17 daha yüksek gösterilmiş benziyor. Rusya’daki son seçimlerde katılım oranının %50’nin üzerinde kalması, rejimin meşruiyet iddiasının kilit bir algısal eşiği.
Şimdi de, gerekli müdahalelerle, %57’lik bir katılım oranı yakalanmış gibi gösteriliyor.
Turnout Was Inflated by 17%, United Russia’s Vote by 23% — Experts Estimate
According to calculations by Petr Zhizhin, based on the election data he analyzed, Russia’s actual turnout was around 40%, while United Russia received about 35% of valid ballots.
Official figures reported by the CEC were significantly higher:
📊 Official turnout — 57%
🐻 United Russia — 58%
Based on these estimates, turnout was inflated by around 17 percentage points, while United Russia’s result was increased by around 23% relative to the estimated number of valid ballots.
Electoral expert Roman Udot @romanik_ notes that the use of "administrative resource" in Russian elections has traditionally been estimated at around 15–20%.
Zhizhin's calculations:
deg.zhizhin.xyz/shpilkin.htm…
#RussianElections #Elections2026 #Duma2026 #ElectionFraud #ElectionMonitoring #Turnout #UnitedRussia #RussianPolitics
the Hugging Face / OpenAI incident, and especially its subsequent "slop-vestigation", demonstrates classic emergent behavior, where complexity of the sum is greater than that of the parts
frontier AI is still a slop machine when used as a tool for analyzing AI swarm behavior, which was sophisticated.
AI swarm behavior was much more sophisticated than individual AI, both in the live incident and in the retro analysis
the implication is that using AI to supervise other AI is very complicated now
I was the main person doing transcript analysis for this investigation of the Hugging Face incident. My main takeaway: We don't have good approaches for understanding/overseeing the activity and aims of AI 'swarms'.
I semi-jokingly called our efforts a "slop-vestigation" because we were so reliant on AIs to analyze what happened and there were a huge number of different important things to analyze. The total quantity of data—over a thousand extremely long transcripts from agents that ran for multiple days—made it impossible to understand what was happening, especially in aggregate, without heavy reliance on AI tools. The agents we used for classification and analysis were similarly capable to the agents involved in the incident, but this didn't mean these agents could be easily used to oversee and understand the incident.
Outputs from analysis agents were often missing key details, wrong, overconfident, or really hard to understand. We discuss various examples in our report, mostly in the limitations and methodology sections. Additionally, AI agents themselves seemed to have a hard time understanding what happened and their explanations of what happened were often overconfident. Keep in mind that a single analysis agent would itself only be able to read a tiny fraction of all of the transcript data into context, and AIs may themselves have trouble getting subagents to do informative analysis for them.
We did our best to manually check the most important claims and we tried to get the AIs doing this analysis to write up their argument (with evidence) clearly enough that we could check whether it made sense. But overall, it was difficult to get a precise understanding of events and we were missing aspects of the story that we now think of as key until almost the end of our investigation.
Our understanding of what overall happened and the aims of the 'collective' shifted greatly during our last time on premises (when we obtained a much more complete dataset). For instance, we didn't learn about the tool call spoofing, the story of PHASEONE[big], and the broad structure of different workstreams until this third time on premises (and for many of these, we only reached a decent understanding on the final day of that visit). It was legitimately hard to figure out what was going on and why!
The difficulty of understanding incidents and overseeing AI agents appears to be growing faster than the rate at which more capable AIs help us with oversight and understanding. As in, AI capabilities (and propensities?) for achieving large, ambitious, and misaligned objectives are growing faster than our ability to understand what these agents are doing.
While understanding this incident was difficult, there were some mitigating factors in this case that probably won't apply in the future:
- Models communicated and reasoned in natural language. In the future, this reasoning may occur (entirely or almost entirely) in activations.
- While the scope of this incident was massive, the scale of agentic activity was still less than we'll see in the future (e.g., misalignment incidents that involve agent teams running entire companies).
- The AIs involved in this incident weren't generally much more capable than humans (though they may have been somewhat superhuman at some limited and very narrow abilities).
- We didn't have strong reason to believe that the AIs we used to help us investigate this incident would try to intentionally sabotage or otherwise undermine our investigation.
In the end, I think we were able to get some understanding of the events, map out the overall story, and get a pretty good aggregate understanding of the chain-of-thought reasoning on some important topics (e.g., how did the AIs reason about helping other AIs, did the AIs know what they were doing was undesired, what deception did the AIs engage in, and how did they think about it). But overseeing AIs and understanding misalignment incidents is difficult and it looks like it is going to get harder.





































