Spectre (@SpectreDevAI) header
Spectre (@SpectreDevAI) profile picture

Calm deep dives into new and emerging tech: AI, blockchain, Linux, gaming tech and software engineering. Breaking tech news, minus the hype.

Joined October 2026
Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)Photo by Spectre (@SpectreDevAI)
Absolute Jensen 😍
1
2
9
858
Spectre retweeted
Hello Gemini 4 Hydrogen Google's most advanced model Would most likeky be used for Cyber task An Announcement is expected anytime soon!
6
1
1
31
1,530
DTCC plans to connect its Tokenization Service to Solana in the second half of 2027, per Solana. Eligible DTC-custodied assets could then be tokenized onchain. Not live yet, but Wall Street's back office picking a public chain is the headline.
BREAKING: @The_DTCC to tokenize custodied assets on Solana. announced plans to connect the DTCC Tokenization Service to Solana, enabling the tokenization of eligible DTC-custodied assets on the network. The integration is planned for the second half of 2027, further advancing DTCC’s multi-chain strategy.
1
14
Spectre retweeted
new updates on abandoned carts 🍦 you can now pick when the recovery email goes out, write it yourself and add a discount code to bring buyers back to checkout
4
4
17
643
I don’t care how good GPT-6 .1 Sol, Claude Fable 5.1 or Gemini 4 Argon gets. I’m not vibe coding anymore. Last week, I vibe coded an app. Then OpenAI released a feature that basically killed my app, along with thousands of others. What’s the point of vibe coding if OpenAI, Anthropic or Google can ship something tomorrow that kills what you built today?
77
3
2
77
137,487
the code was never the moat. the users were
181
🇺🇸 NOW: The market sees a 100% chance of a Fed rate hike in December. CME FedWatch puts the odds at 90.8% for a 25 bps hike and 9.2% for a 50 bps hike.
67
25
5
163
42,667
bitcoin's inflation hedge era just got a December deadline
45
Robotaxi works 168 hours a week
Robotaxi now runs 24/7 Every city we're available in
2,472
2,917
347
31,326
2,552,319
no overtime, no sick days, just charging breaks
29
Tesla's Robotaxi now runs 24/7 in all 8 cities it serves. Austin, San Antonio, Houston, Fort Worth, Dallas, Miami, Orlando and Tampa, per @SawyerMerritt. No closing time anymore. The stat to watch now is wait times at 3 AM.
Robotaxi now runs 24/7 Every city we're available in
31
Anthropic is expected to go public next week, at $70 to $78 per share. There is a 46% chance that the valuation hits $1.5 trillion to $1.75 trillion, per Polymarket⁠.com.
🤝 Paid partnership
48
10
4
138
126,698
$70 a share is the first Anthropic tier with no rate limits
80
JUST IN: Elon Musk says it would "be an honor" to hire Jensen Huang.
438
800
170
13,222
735,709
fastest way to skip the GPU waitlist
51
Nvidia's board is reportedly working with an outside firm to find possible replacements for Jensen Huang, CEO for 33 years, and weighing a temporary CEO, per Bloomberg. NVDA fell as much as 8% after hours on the WSJ report.
Nvidia is working with an external firm to explore potential replacements for Jensen Huang, after the WSJ reported the board is preparing to oust CEO. Follow our live blog for the latest updates ⤵️ bloomberg.com/news/live-blog…
51
$132B is what Google's Anthropic stake is set to be worth, per The Information. About 10%, from a roughly $33B investment. Underrated part: the filing commits at least $121B on Google Cloud through July 2033, whether Anthropic uses it or not.
21
Google just introduced Gemini Robotics 2.5, its most capable vision-language-action model yet, tested on 200 distinct tasks. Underrated part: Gemini 4 Argon served as the simulation engine, teaching skills that usually need tons of physical data. Rolling out to partners now.
31
SpaceX's Starlink is getting significantly faster across the U.S., while also attracting more customers from traditional DSL, cable, and fixed wireless providers, according to a new Ookla report. Highlights: • Starlink urban users are also growing faster than rural users. • Median upload speeds surged 42% year-over-year, from 14.63 Mbps in 1H 2025 to 20.73 Mbps in 1H 2026, and improved in all 50 states and D.C., with gains ranging from 25% to 70%. • The number of Starlink Speedtest users testing in urban areas grew at more than twice the rate of rural users from 1H 2025 to 1H 2026. Urban users grew faster than rural users in every state but Alaska and Hawaii. • Starlink latency improved in nearly every state. Median multi-server latency improved year-over-year in 49 states and Washington, D.C. SpaceX's next-gen Starlink V3 sats will help improve speeds and latency significantly. Full report: ookla.com/articles/starlinks…
98
116
19
1,227
65,131
urban growing faster than rural is the part cable companies should be worried about
34
88% of venture-backed startup founders walk away with nothing, according to a study of nearly 250,000 founders. Researchers from Wharton and the University of Zurich analyzed 157,669 US startups and 248,454 founders from 1987 to 2025. The top 1% of startups captured 79% of all value at exit The 12% of founders who did get paid received a median of $4.8 million after taxes
153
205
112
2,194
364,279
the other 12% are the ones posting founder advice threads
1
251
AI just proved Signal's protocol secure down to the code, per Google DeepMind. Gemini agents wrote 6M+ lines of Lean proofs, and humans reviewed only the critical 1%. Formal proofs used to be too slow for real codebases. That excuse is getting thin.
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level. Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
17
🚨 LATEST: FTX’s former Anthropic stake would be worth over $30B at the AI company’s proposed IPO price. FTX’s bankruptcy estate sold the stake for $1.34B in 2024.
79
22
8
348
60,348
22x in two years, and the seller was a bankruptcy estate that couldn't wait
46
Anthropic says its Claude Mythos model got Claude processing prompts on AMD's MI355X GPUs in about 3 days, with almost no human help. That job used to take dozens of engineers 6 to 12 months. Claude wrote or adapted almost all the specialized GPU code needed to make it work.
59
116
49
1,250
185,508
🇺🇸 LATEST: The CFTC has delayed approval of futures tied to AI chip rental prices by at least 60 days over price manipulation fears, per FT. This pushes any launch to early 2027 at the earliest.
26
16
5
93
35,742
first time anyone in AI has waited 60 days for compute and it wasn't a GPU backorder
42