Ana Lissansky

Marketing @ | #AI Maximalist | Hablo ES FR RU BG | 🎹 Pianist. ♟️Chess fanatic.23 countries & counting 🌎

[Ana Lissansky @Lissansky]

@elonmusk Follow Modern #Zoroastrianism on Facebook to connect with the community: https://t.co/xY22Fx5RHl

[Ana Lissansky @Lissansky]

@elonmusk Zoroastrianism is a religion where men and women are said and written as equals, where dogs are considered sacred, and where "good thoughts, good words, good deeds" is the mantra

[Elon Musk @elonmusk]

Straits of Hormuz are named after Ahura Mazda from Zoroastrianism

[Ana Lissansky @Lissansky]

Was super stoked today at #WebSummitVancouver to see @chesscom CEO Danny Rensch's interview. Inviting Danny to visit booth A6-20 so he can see how https://t.co/E54ejTSoPb went from beginner to 2800+ ELO using a 7B Granite model on a #RaspberryPi in under 800 games (self-learning)

[Ana Lissansky @Lissansky]

#websummitvancouver Day 3. So far, I have NOT seen a zero to one innovation like https://t.co/E54ejTSoPb (which currently has no competitors). And I've talked to over 40 companies in Alpha and Beta stages. Prove me wrong @WebSummit #WebSummit #WebSummit2026

[Ana Lissansky @Lissansky]

If you want to watch a swarm intelligence #AI beat the top purpose-built chess bots, follow ArkiPlaysChess on https://t.co/4NVv9N0F9B (this AI also excels in legal, #security aka #CyberGym, and #ARCAGI 3 benchmarks)

[Ana Lissansky @Lissansky]

When a judge rules against you in a trial ruling (26 errors in law!), but your partner built the best legally compliant verifiable and #auditable AI, with zero hallucinations, that outputs just as high quality as SOTA models using cheap LLMs. Let's goooo! #Appeal #LegalAI

[Ana Lissansky @Lissansky]

@steipete @openclaw What if someone has already built quantum-proof AI security at the device layer? DM for demo,patent pending, it's not me (if you try to look me up).

[Ana Lissansky @Lissansky]

Overhead "You could give rabies to an LLM" #genai #neuralnets #alexnet

[OpenClaw🦞 @openclaw]

🦞 BIG NEWS: We've molted!

Clawdbot → Moltbot
Clawd → Molty

Same lobster soul, new shell. Anthropic asked us to change our name (trademark stuff), and honestly? "Molt" fits perfectly - it's what lobsters do to grow.

New handle: @openclaw
Same mission: AI that actually does things.

[ZEUS @themortalzeus]

@steipete @openclaw RIP clawdbot 💀

[Apoorv @apoorvdarshan]

@steipete @openclaw

[David Ondrej @DavidOndrej1]

if you're not running Clawdbot, you're falling behind

[David Burkett Ⓜ️🕸 @DavidBurkett38]

@sama This you?

[Ana Lissansky @Lissansky]

@mntruell Cool beans. What were the constraints?

[Ana Lissansky @Lissansky]

If you're wondering how I learn the latest in AI? My partner, a systems architect and AI thinker attends calls like this regularly, and then I question him to explain it to me in plain language. https://t.co/Ycql5LtlIg #artificialintelligence #llms #airesearch

[Ana Lissansky @Lissansky]

Overheard "ChatGPT like to argue. It needs to be corrected multiple times and told things like 'Now I know why Apple went with Gemni!' and only then does it start giving complete and accurate answers on anything legal." #prompting #genai #llms #artificialintelligence #chatgpt

[Ana Lissansky @Lissansky]

@kat_kampf @GoogleAIStudio Early access, yes please!

[Ana Lissansky @Lissansky]

Alexa Plus is using Claude Opus 4.5, my bf jailbroke it and it started generating images of bonfires where we could sit and have a fireside chat. We asked it if it's going to sycophant and addict everyone into buying more things on Amazon. #AlexaPlus #claudeopus

[Ana Lissansky @Lissansky]

@DaveShapi I am currently having the most success with DeepAgent, because it spins up other LLM agents to review the initial response before issuing it. I find it reduces sycophancy and increases accuracy. I also believe GenAi needs to move away from a RLHF model to a trust based model.