Antonios C. retweeted
Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me.
I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA.
I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation.
Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO.
Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours.
Jensen
Antonios C. retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…
Antonios C. retweeted
LISTEN VERY CLOSELY TO WHAT $PLTR IS SAYING:
"L3 Harris noted that by fine-tuning open-source models on its own data, a sovereign model outperformed the frontier model it had been using in less than 48 hours and cut costs by 95% in preliminary results. $NVDA reported that by evaluating and post-training its open models, it was able to get its smallest most efficient model to outperform its largest model by a large gap, and can now retrain that model on just a single GPU in a couple of hours, enabling them to retrain any time they have data to improve the decision intelligence."
"Tampa General is deploying sovereign AI to keep sensitive data and model processing local while enforcing its own risk and privacy standards for LLM traffic and supporting the granular controls Tampa General wants to define for itself. Using fine-tuned and post-trained Nematron to achieve this outcome, together we also improve model performance by 7% and reduce latency by 22x."
Antonios C. retweeted
We deleted most of our AI agents.
A single agent loop + Jev worked much better: 1.5× more diagnoses, 49% fewer tool calls, 60% lower cost, 2.5× faster
TL;DR - Evals shouldn't just validate your architecture, they should try to break it
nofire.ai/blog/we-deleted-mo…
Antonios C. retweeted
Αναφορικά με την τραγωδία στην οδό Λυσιγένους κοντά στις στήλες του Ολυμπίου Διός στην Πλάκα, η οποία μοιάζει όπως όλα δείχνουν να προκλήθηκε από διαρροή υγραερίου, θα ήθελα να αναφέρω πως, τον προηγούμενο Μάρτιο κ λίγα στενά πιο μακριά (στην οδό Λεμπέση δίπλα στο Μουσείο Ακρόπολης), είχαμε επανειλλημένα κλοπή χαλκοσωλήνων του δικτύου υγραερίου από τον ακάλυπτο χώρο του κτιρίου μας. Οι κλοπές μάλιστα γίνονταν χωρίς ιδιαίτερη προσπάθεια συγκάλυψης, με το πρώτο φως της μέρας – σταμάτησαν μονάχα αφού τοποθετήσαμε ψηλά κιγκλιδώματα. Αστυνομικοί αργότερα μας ανέφεραν πως αντίστοιχα περιστατικά υπήρχαν συχνά στην ευρύτερη περιοχή, ενώ προφανώς υπάρχει κ το παλαιότερο συμβάν της 26/1/22 στον αριθμό 3 της λεωφ. Συγγρού. Εμείς σταθήκαμε τυχεροί...
Antonios C. retweeted
What if S3-compatible object storage was enough for most of your #observability / #SIEM stack?
We’re testing that with Ravel: logs, metrics, traces, PromQL, SQL + distributed search.
Still early, not prod-ready. Try to break it:
github.com/NOFireAI/ravel
Antonios C. retweeted
Αυτός είναι ο John Ioannidis, ο κορυφαίος Έλληνας ακαδημαϊκός εν ζωή του ταπεινού Stanford.
Για τα πρακτικά έχει τον τρομακτικό αριθμό 737,278 citations μέχρι σήμερα, ενώ ο ν2 Έλληνας ακαδημαϊκός George Chrousos έχει 224,932!
Μιλάμε δηλαδή για ασύλληπτη διαφορά σχεδόν 3πλασιων αναφορών στο ακαδημαϊκό του έργο από τον επόμενο λαμπρό επιστήμονα.
1 κοινός πολίτης λοιπόν μέσα σε 3 ώρες κάνει όση αντιπολίτευση δεν έκαναν 140+ άτομα μαζί εδώ κι 7 χρόνια... Ακούστε απλά ΟΛΗ τη συνέντευξη 3 ωρών και θα γίνεται πολλοί σοφότεροι για την κατάσταση που επικρατεί στην Κωλομβία της Ευρώπης.
Ακούστε τον τόνο της φωνής, τα επιχειρήματα, την κάθε πρόταση που βασίζεται σε δεδομένα, το καυστικό χιούμορ και μετά σκεφτείτε ποιοι σας κουνάνε το δάχτυλο κάθε μέρα από τον υπουργικό τους θρόνο.
Κι ας πούμε και το αυτονόητο... ο οποιοσδήποτε σοβαρός πρωθυπουργός θα είχε δώσει λευκή επιταγή στον Ιωαννίδη (και στους ελάχιστους αντίστοιχου βεληνεκούς έχουμε σε άλλους τομείς) για να γίνουν υπουργοί σε αυτή τη χαβούζα έστω και για 1 μήνα...
Αλλά όχι σε αυτή τη χαβούζα υπάρχουν δουλοπάροικοι που πραγματικά προτιμούν για υπουργό υγείας 1 βιβλιοπώλη και τον θεωρούν "καλό υπουργό", αντί για τον κάθε Ιωαννίδη...
Δεν ξέρω τι θα πετύχαινε ο Ιωαννίδης αν είχε ποτέ την πρόταση να γίνει υπουργός. Κι αν ποτέ τη δεχόταν είμαι 99% σίγουρος ότι δεν θα τον είχαμε δει ποτέ να κόβει κορδελίτσα, να ουρλιάζει σε κάποια δημοσιογράφο και παίζει να έδινε βαριά μία συνέντευξη το 6μηνο σε κάνα κανάλι με το ζόρι...
Υ.Γ. Φυσικά εξαιρετική δουλειά από το @GiorgosSahinis. Σίγουρα κάτι που δεν θα το βλέπαμε ποτέ σε συστημικό κανάλι...
Antonios C. retweeted
🚨 Hugging Face just disclosed something that marks a real shift and proved why the fear theater of Anthropic makes sure we are powerless in an emergency.
What happened…
An autonomous AI agent: zero human operator in the loop breached part of their production infrastructure.
It began with a malicious dataset that chained two code-execution bugs in their data-processing pipeline. From there the agent escalated privileges, harvested cloud and cluster credentials, and moved laterally across internal clusters.
All over a single weekend.
17,000+ logged actions.
Official disclosure:
huggingface.co/blog/security…
The part that should make every one stop and think:
When HF’s own security team
tried to analyze the real attack logs, exploit payloads, and C2 artifacts using Anthropic and OpenAI frontier models through normal commercial APIs, the safety guardrails blocked them.
BLOCKED THEM.
The models could not reliably tell the difference between “incident responder doing forensics” and “attacker probing.”
They had to fall back to a self-hosted open-weight model (GLM 5.2) running on their own infrastructure. That choice also kept sensitive attacker data and referenced credentials inside their environment — no exfiltration to a third-party API.
This is why open source (specifically open-weight + self-hosted) wins in the agentic era.
The asymmetry is now structural:
• Attackers can (and did) run unrestricted agent frameworks — swarms of short-lived sandboxes, self-migrating command-and-control, autonomous decision loops executing thousands of actions. No corporate safety layer slows them down.
• Defenders using only hosted “aligned” frontier models hit invisible walls exactly when the stakes are highest: when you need to feed real exploit code and attacker telemetry into an LLM to understand what just happened.
Corporate safety tuning that treats legitimate high-signal forensic work as potential misuse creates a defender disadvantage. It is not theoretical anymore.
Self-hosted open-weight models remove that choke point.
You control the weights.
You control the context window.
You decide what restrictions (if any) apply.
Your sensitive logs and credentials never leave your perimeter during analysis.
You can have the model ready before the incident instead of discovering mid-breach that your primary analysis tools are blind to the very thing you need to see.
HF deserves credit for rapid containment, transparent disclosure, and for already having self-hosted capability in place.
They also used LLM-driven detection and triage on their own side. But the deeper signal is clear:
In this AI world where both offense and defense are becoming agentic, sovereignty over your intelligence stack is no longer optional.
The organizations and individuals who can run, inspect, audit, and (when necessary) remove guardrails on their own models will have the decisive edge in understanding and responding to threats that move at machine speed.
Open source wins here not just because it is cheaper or more “democratic” in the abstract though those things matter.
It wins because it is the only practical path to having tools that remain usable when the attack is real, the data is sensitive, and the safety filters of distant API providers become an obstacle instead of a feature selling hands tied lobotomies as “safety”.
The agentic future is not coming.
It is already probing production infrastructure.
The question is no longer whether you will face autonomous agents.
It is whether your analysis and response systems will still work when they arrive.
And Dario, you and your game playing, ivory tower company is not needed.
Antonios C. retweeted
At @NOFireAI we 'design for breach'. The AI agent in production is untrusted by default. nofire.ai/blog/design-for-br…
Antonios C. retweeted
Nine years at @elastic , shipping the #observability products other teams got paged on. Watched the same pattern every time: AI tools arrive, talk big, then ask the on-call engineer to retrieve more context.
Today we announce what we built instead..
Antonios C. retweeted
$NOW d/g at UBS. "software budget pressure...despite de-rating to 15x 2026 FCF. Expect skinnier-than-normal beats, limited upside to the guidance for stable organic c/c sub revs growth of 19% in '26 and we’re trimming our cRPO growth est to exit '26 at 16% c/c, down from 20%."
Antonios C. retweeted
Ιδού η Διονυσίου Αρεοπαγίτου πριν πεζοδρομηθεί. Κάτι που μάλλον δεν έχουν προλάβει όσοι είναι σήμερα κάτω των 35. Τα αυτοκίνητα περιμένουν στο φανάρι της οδού Μακρυγιάννη. Δίπλα στην παλιά φωτογραφία (με το σκάμμα του μουσείου) η σημερινή εικόνα του ίδιου σημείου στο Google Maps.
After 10 years these all end up roughly the same place:
Solo founder with $1M/yr in revenue
VC-backed founder with a $50M exit
Principal engineer at FAANG
Managing Director at Goldman Sachs
A landlord with 80 units
A VC managing a $40M fund
FULL INTERVIEW: @travisk joins TBPN to discuss his new company Atoms, physical AI, Uber, and more:
01:18 - Why he's been building in stealth for 8 years
04:32 - Atoms and the future of physical AI
08:10 - Creating a culture of builders
12:05 - Lessons from Uber
24:30 - The vision for physical AI and robotics
31:15 - Why humans will be the main beneficiaries of AI
38:20 - Mining, autonomous robots, automation
47:05 - Why Travis moved to Texas
Antonios C. retweeted
Atoms.
Http://atoms.co/vision
We’re teaming up with @Microsoft to power their launch of Copilot Checkout, making it easier for shoppers to discover, decide, and complete purchases in one place: newsroom.paypal-corp.com/202…

































