Deshayes Yann

Senior sofware engineer @runneth. Previoulsy: @Behaviour Lead Front-end, @platformsh, @hellojam_. guitare/bass player, martial art.

[Deshayes Yann @_Smi13]

@shanebarakat Software engineers tell the AI what to do, vibe coder tell the AI what they want. Script kiddies is not a new concept.

[Chinmay @_chinmaymk]

Introducing cribl-decision-1.0!

Every telemetry event is a decision: where to route it, how to enrich it, and what needs attention. Today, we're releasing a 4B parameter, open-weight model that returns probabilities for your supplied choices in a single forward pass - a foundation for bringing AI decisions into telemetry pipelines.

The weights are available on huggingface: https://t.co/ADrNchkx1i

Read more: https://t.co/eHvFIaF3PR

[Mike Choi @mikechoi]

MCP events might be the most underrated ship so far

Today your agent only wakes up when you message it or put it on a schedule

We got chatgpt to hand a task to runneth, sleep, and get woken by an event when it was done

No polling, no copy pasting!! https://t.co/EJwre6NXac

[Fireworks @FireworksAI_HQ]

GLM-5.3-Flash: GA on Fireworks.
We held it two days over a reasoning-length gap we couldn't explain. Same weights ≠ same model. Quality first.
Start building → https://t.co/yHRs5FOUOQ

[kwindla @kwindla]

Introducing PhoneLLM, an open model for voice agents.

GPT 5.6 Terra performance on typical voice agent tasks at 1/3 the latency and 1/18 the cost.

For voice agents, we need models that are both very low latency and very good at tool calling and instruction following.

There's a trade-off here, and we often have to compromise on either latency or capability when building voice agents. With PhoneLLM (and the training and data stack that made this model possible) we're fixing this problem.

For the last couple of years, most of the effort in frontier model development has gone towards leveraging test-time compute. Which is awesome! Models of all shapes and sizes are available that perform really, really well ... if you have "thinking" turned on for your model.

But if you need your agent to respond at voice conversation speed, you can't use thinking models.

PhoneLLM is a full-weights fine-tune of NVIDIA Nemotron Nano 30B. We trained on a wide range of real-world telephone and customer support use cases. The training focused on taking the excellent Nano 30B base capabilities and teaching the model to do typical voice agent tasks with thinking disabled.

The results are really good: accurate tool calling and concise, on-topic responses in long conversations.

And fast: TTFAT measured server-side is <100ms if you run PhoneLLM on a lightly loaded B200. :-)

But seriously, when we characterize model latency, we do it with full, end-to-end, batched request simulations using real Pipecat voice agent pipelines.

You can serve more than 80 concurrent agents on a single B200 with P95 end-to-end TTFAT <600ms. Including network overhead. That's an LLM cost-per-minute around $0.0025. (1/4 of a cent.) At a latency lower than any third-party API offers today.

More details about this model, including weights on @huggingface, how to spin it up with one click on Modal, and a starter project repo you can clone, are in the thread ...

[Z.ai @Zai_org]

Introducing GLM-5.3-Flash

- Leading capabilities at a highly competitive price
- Natively multimodal with a 1M-token context window
- A 320B-A18B model released under the MIT License
- Previously previewed as Ox Alpha, running entirely on Chinese AI chips

Blog: https://t.co/tzOmB7gdZP

Available now across all official platforms:

Weights: https://t.co/9LRMahY9Wa
API: https://t.co/VcaQnzYmS9
Coding Plan: https://t.co/Nk8Y98HNhU
ZCode: https://t.co/Peepqv4XSx
Chat: https://t.co/WCqWT0qCQb
AutoClaw: https://t.co/aGEG5HqTTb

[Deshayes Yann @_Smi13]

@GoodnotesApp Hello! we are trying to change our siubscription to be able to have the app on windows AND iOS, but we are stuck in the android and windows only subscription. Can you help please? I sent you support messages already, no response...

[Deshayes Yann @_Smi13]

@LinkedInHelp Hello, thx for the response.I sent it in dm

[Deshayes Yann @_Smi13]

@LinkedInHelp hello, my account is locked, I tried to DM this account but 2 weeks without any response. Can I have some help please ?

[Deshayes Yann @_Smi13]

@vfsglobalcare I sent the link via PM.

[Deshayes Yann @_Smi13]

@vfsglobalcare Hello, I'm trying to confirm the appointement I have tomorrow in Lyon for a Canadian visa but I have this error, can you help me please?

[Deshayes Yann @_Smi13]

@vfsglobalcare I tried on windows as well, I’m in incognito mode, the server return a 500. I can see that le code in the query string is not encoded properly, do you think it is possible to generate an other link?

[Deshayes Yann @_Smi13]

@EIC_IEC Bonjour, j'ai fait une demande pour un permis jeune professionnel avec une entreprise à Montréal le 3 février 2022 depuis le Canada. On m'a demandé de payer les frais biométrique le 11 mars 2022, ce que j'ai fait. Je n'ai à ce jours pas de nouvelle, pouvez-vous m'aider ?

[Deshayes Yann @_Smi13]

@EIC_IEC Bonjour, si ma femme a une offre d’emploi et obtient son pvt, pourrais-je la suivre ? Si oui avec quel visa ? Car d’après ce que j’ai compris, le pvt est individuel et au vu des conditions d’entrée actuelles, je ne peux même pas venir en visa touriste ? Merci

[Deshayes Yann @_Smi13]

@EIC_IEC Bonjour, si ma femme a une offre d’emploi et obtient son pvt, pourrais-je la suivre ? Si oui avec quel visa ? Car d’après ce que j’ai compris, le pvt est individuel et au vu des conditions d’entrée actuelles, je ne peux même pas venir en visa touriste ? Merci

[Deshayes Yann @_Smi13]

@DasSurma Google spreadsheet yes

[React Training @ReactTraining]

📣 Just published: React Router v6.0.0-beta.0 👀
- We're in beta now! That means we are closer to stable, so fewer breaking changes from here on out 🤘
- Lots more docs, including a full API reference 📖
- Better sourcemaps 🗺
Release notes 👇
https://t.co/srRGqxtPaf

[Deshayes Yann @_Smi13]

@Sardoche_Lol Hey mec qu’est-ce qui se passe là ? Appelle si tu as besoin.

[@Pwn20wnd @Pwn20wnd]

#unc0ver v5.0.0 is NOW OUT.