Boris Wang | Founder @ jenova.ai

interesting phenomona i saw in past few years of digital nomading -- polygamy is still very much alive and might be resurging

@9mmsmg why is onlyfans responsible for retarded men spending money for online girls lol

Onlyfans is responsible for introducing thousands and thousands of young women into online sex work. It is responsible for the destruction of marriages of men and in many cases the financial ruin of families with men spending thousands on girls they'd never meet. Bury the site with it's founder.

@wanggege911fbi @JohnLone246312 要客观看历史 清朝正面贡献：东北，内蒙，新疆，西藏 清朝负面贡献：其他所有方面，尤其文化上

@boriswang01 @JohnLone246312 看看新中国怎么定义的 你现在学的是改开后的理论 被人掺了沙子 新中国的领土跟清朝没关系 那些领土 都是历朝历代有历史依据的 倒是清朝 直接将中国地图缩水 丢失东南亚领土是最大的罪过

是這樣嗎？

@wanggege911fbi @JohnLone246312 提出一个除了领土外的满清贡献

@boriswang01 @JohnLone246312 怪不得叫清华了 清除华夏历史

@sayinshallah jesus himself was a jew

The world is finally seeing why Jesus even lost his composure around the jews

@HarryStebbings well the ARR bar for series A now is also much higher so its all in sync

I miss the days of a $15M on $75M Series A. Can’t do anything for less than $30M on $160M today.

about time

@ci00759584 its good that you didn't include qing, would have drastically lowered the aesthetic quality of this video

一镜到底穿越三千年！

@chatgpt21 and its all shit.

I don’t think people are appreciative enough that OpenAI really has been doing a monthly update GPT-5.1 — November 2025 GPT-5.2 — December 11, 2025 GPT-5.3 (Codex / series) — February 5, 2026 GPT-5.3 Instant — March 2026 And now GPT 5.4 this week. And keep in mind these weren’t small jumps on benchmarks they were meaningful. At this rate GPT-5.5 — April to May 2026 GPT-5.6 — May to July 2026 Then the major jump: GPT-6 — late 2026

@iruletheworldmo no model has had this much anticipation since GPT-5, hope it doesnt disappoint

damn. i told you. just deepseek to go. 5.3 instant is an incredible model fyi, give it a try.

Google @GoogleDeepMind facing compute crunch recently, both AI Studio or their enterprise Vertex AI the rate of model unavailability/instability has been unhealthily high these few days. Model providers really need to start thinking about cutting usage to consumer user and allocating more compute to API users.

@Tablesalt13 one hour is sufficient tbh, i am canadian

I am now 100% confident that the USA could take Canada in 2 hours I have no doubt at this point.

@Pauliespasta @iruletheworldmo o right i totally just forgot about atlas😂😂😂 damn thats bad

@boriswang01 @iruletheworldmo dont forget the Atlas thing! who the fuck asked for it? Meanwhile i use all of claude stuff daily

anthropic have now shipped openclaw, cowork (also for windows) dario and the gang are unfathomably ahead and their velocity is insane. it’s a fair observation that they may have reached escape velocity. crushing every benchmark, eating enterprise, department of war begging, taking over openai revenue. this is what i’d imagine would happen if one of the labs made a huge breakthrough.

@JenovaAIAgent @iruletheworldmo

Introducing the latest results of our Long-Context Agentic Orchestration Benchmark. • 31 high-complexity, non-coding scenarios (100k+ tokens) where the model must select the correct next-step action using proprietary orchestration logic with no public precedent — a pure test of instruction following and long-context decision-making. • All models run at minimum thinking/reasoning settings and temperature 0 — simulating production orchestration where determinism and speed are critical. • Claude and Gemini dominate. Chinese open-source models underperform. GPT-5.2 struggles without extended reasoning. Full methodology and analysis in the comments. @AnthropicAI @GoogleDeepMind @JeffDean @OfficialLoganK

@aakashgupta where are you pulling the information? perplexity never embeded ads in their AI responses directly

Perplexity just admitted the Google ad model can’t work in AI search. And the math explains why. Perplexity hit ~$200M ARR by late 2025 on a $20B valuation. Their 2026 revenue target is $656M. That implies 230% growth, almost entirely from subscriptions and enterprise contracts, because ads generated so little revenue they weren’t worth the trust damage. Here’s what happened behind the scenes. Their head of ad sales, Taz Patel, left in August 2025. By October, they stopped accepting new advertisers entirely. The internal calculus was simple: brands didn’t want to buy ads in conversational AI because their messaging gets rewritten, condensed, or misattributed. Traditional search ads sit alongside results. AI ads get absorbed into the answer. There’s no clean separation, and advertisers hated it. The economics told the same story. At 60M MAUs, Perplexity’s ad inventory is a rounding error compared to Google’s 8.5 billion daily searches. You can’t build an ad business at that scale because CPMs in AI search have no pricing benchmark, no established measurement framework, and no reliable attribution model. Advertisers had no idea what they were buying. This tells you something about the entire AI search category. OpenAI is moving the opposite direction, introducing ads to ChatGPT’s free tier this month. But OpenAI has hundreds of millions of users and Microsoft’s ad infrastructure behind it. Perplexity has 60M MAUs and a five-person enterprise sales team. The math only works at Google-scale volume, and no AI startup has that. So Perplexity is doing the rational thing: targeting CEOs, doctors, and finance professionals at $20-$200/month, because 15M paying users at high ARPU beats 60M free users generating pennies in ad revenue. Anthropic reached the same conclusion without even trying ads. Claude went straight to enterprise. The real signal here is what Perplexity’s CEO said a year ago versus what executives are saying now. Srinivas publicly predicted ads would become a cornerstone of revenue. Now his team says they might “never ever need to do ads.” That’s a full 180 in under 12 months, which means the internal data was bad enough to kill the entire strategy. Every AI search startup pitching an ad model to VCs should update their deck.

hey @JeffDean @OfficialLoganK @OmkarSavant_ why not allow the Pro models to have a minimal/none thinking mode? consumers dont want variable response latency ranging from 10 - 60 seconds. this is forcing us to use Anthropic models when we want to use Gemini