sylvain.peyronnet.eu
I live by and for the algorithms.
Currently CEO at @babbartech (also @yourtextguru)
80 posts
467 followers
299 following
Regular Contributor
Active Commenter
comment in response to
post
You forgot
1.5/ Design your own algorithms so that your tool truly has added value
Otherwise it's just "more money, less time", not worth it IMO.
comment in response to
post
But has anyone ever seen Tory Gray and Lily Ray in the same room?
comment in response to
post
Frontpage FTW!
comment in response to
post
I'm late!
comment in response to
post
This one is a good read ->
comment in response to
post
Crawling the web full scale is not always easy, I wonder why AI operators haven’t yet approached full scale crawl operators 🤷♂️ they’re not hard to find: the cloudflare radar gives their names ^^
comment in response to
post
Ah, at first I saw 'finally bearded'
(Sorry, I couldn't help myself 🤪)
comment in response to
post
Probably the best thing that came out of the French political world in the last months (years?)
comment in response to
post
That is an excellent idea! A bientôt alors !
comment in response to
post
Not my paper ;) (Though I would have liked to)
The accuracy of info is an issue, safeguards need to be put in place (using graphRAG, for ex).
Our collective mistake is to assume that speech inherently contains knowledge. There’s no reason for that to be true (otherwise, we'd all be very wise 😅).
comment in response to
post
This is all new and somewhat uncharted territory, so adjustments will be made in the coming weeks to continually improve quality.
Additionally, with searchGPT guides, there's no simple notion of competitors or SERPs, so naturally, no "SERP competitors" tab in the tool for these guides.
comment in response to
post
On the other hand, SearchGPT is a "secondary" engine: an informational need is communicated in natural language to the LLM, which handles its own research.
This more complex workflow shifts from keywords to understanding search intent. Complex queries yield better responses.
comment in response to
post
Google and Bing are "primary" search engines, keyword-focused, running their own QBST for initial content filtering. Moving from Google to Bing was just about data acquisition for us. Both engines handle semantics slightly differently—marginally, but it's noticeable.
comment in response to
post
Il y a du ML à tous les niveaux, et le module qui s'occupe des snippets (snippetbrain) peut réecrire titre et description pour maximiser le click through rate de ton snippet.
Parfois ça déconne complètement ^^
comment in response to
post
En octobre 2023 déjà tu crois ?
comment in response to
post
Thank you, looks very interesting. Is there an API/white-label option available?
comment in response to
post
comment in response to
post
There is already an academic paper about GEO: Generative Engine Optimization 😅
Bonus: a link to the paper -> arxiv.org/pdf/2311.09735
comment in response to
post
En premier le manque de réflexivité je dirais. En 2017 quand j'avais écrit sur mon abandon de mon poste de PU j'avais mis "La pauvreté intellectuelle" dans les points qui me gênait
une-belle-etoile.fr/quelques-bon...
comment in response to
post
L'Etat ne risque jamais la requalification de ses vacataires micro-entrepreneurs en salariés ou bien personne n'a jamais tenté le coup pour avoir un poste ?
comment in response to
post
C'est pas pour faire un rappel K200/IA ? Je ne suis pas sur que ce soit plus glorieux, mais il y a qqun qui a eu une diée au moins 🤷♂️