Priyaa

782 posts

Priyaa banner
Priyaa

Priyaa

@pritopian

Building graphic design AI models that actually listen to you. Founder @world_lica. Fellow @southpkcommons. Ex-@waymo, @snapchat, @microsoft. Lurking on X.

San Francisco Katılım Nisan 2023
799 Takip Edilen2.8K Takipçiler
Priyaa
Priyaa@pritopian·
@thesayannayak Why find users when you can make your own and rebrand them as AI agents?
English
0
0
0
11
Sayan
Sayan@thesayannayak·
At this rate everyone’s gonna have their own app and zero users.
English
561
647
10.5K
532.4K
Priyaa
Priyaa@pritopian·
@imjaredz Dalida in presidio?! The best.
Español
1
0
1
16
Jared Zoneraich
Jared Zoneraich@imjaredz·
How do you stand out in the age of agents? Now that every website has cool animations. Now that every meal is a bowl. ==the antidote is soul== The best meals are made with love. The best garmentos obsess over button details. Good Design is taste. It conveys a message. It’s human. We’ve redesigned promptlayer.com and we’re really proud of the result. It’s inspired by my favorite restaurant in SF. Every icon is hand drawn.
English
16
5
45
2.2K
Sumanth
Sumanth@sumanthd17·
Just save yourself 4hrs and watch it on netflix
English
5
0
11
1.6K
brett goldstein
brett goldstein@thatguybg·
it’s honestly uncanny how accurate this is: - ravenclaw: Anthropic - gryffindor: OpenAI - slytherin: X AI - hufflepuff: Google/Gemini
English
29
8
246
26.6K
Lenny Rachitsky
Lenny Rachitsky@lennysan·
Even though every AI company is building their own version of OpenClaw (which is smart!), I haven't seen any of them get anywhere near the love and passion that OpenClaw inspires. There's something special about the OpenClaw experience that's hard to copy.
Thariq@trq212

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

English
53
11
195
26.8K
Priyaa
Priyaa@pritopian·
@anshublog SOC2 is really the way to clear enterprise procurement queues for startups 🥲 so for more early stage companies these quick compliance certificates can sound compelling.
English
0
0
3
274
Anshu Sharma 🌶
Anshu Sharma 🌶@anshublog·
Every great data security breach happened at a company that was SOC-2 compliant. Let’s delve deeper: - Equifax SOC-2 ✔️ - T-Mobile SOC-2 ✔️ - Ticketmaster SOC-2 ✔️ - Change Healthcare SOC-2 ✔️ Pick a good SOC-2 vendor but even the best don’t really give you any security.
Ryan@ohryansbelt

Delve, a YC-backed compliance startup that raised $32 million, has been accused of systematically faking SOC 2, ISO 27001, HIPAA, and GDPR compliance reports for hundreds of clients. According to a detailed Substack investigation by DeepDelver, a leaked Google spreadsheet containing links to hundreds of confidential draft audit reports revealed that Delve generates auditor conclusions before any auditor reviews evidence, uses the same template across 99.8% of reports, and relies on Indian certification mills operating through empty US shells instead of the "US-based CPA firms" they advertise. Here's the breakdown: > 493 out of 494 leaked SOC 2 reports allegedly contain identical boilerplate text, including the same grammatical errors and nonsensical sentences, with only a company name, logo, org chart, and signature swapped in > Auditor conclusions and test procedures are reportedly pre-written in draft reports before clients even provide their company description, which would violate AICPA independence rules requiring auditors to independently design tests and form conclusions > All 259 Type II reports claim zero security incidents, zero personnel changes, zero customer terminations, and zero cyber incidents during the observation period, with identical "unable to test" conclusions across every client > Delve's "US-based auditors" are actually Accorp and Gradient, described as Indian certification mills operating through US shell entities. 99%+ of clients reportedly went through one of these two firms over the past 6 months > The platform allegedly publishes fully populated trust pages claiming vulnerability scanning, pentesting, and data recovery simulations before any compliance work has been done > Delve pre-fabricates board meeting minutes, risk assessments, security incident simulations, and employee evidence that clients can adopt with a single click, according to the author > Most "integrations" are just containers for manual screenshots with no actual API connections. The author describes the platform as a "SOC 2 template pack with a thin SaaS wrapper" > When the leak was exposed, CEO Karun Kaushik emailed clients calling the allegations "falsified claims" from an "AI-generated email" and stated no sensitive data was accessed, while the reports themselves contained private signatures and confidential architecture diagrams > Companies relying on these reports could face criminal liability under HIPAA and fines up to 4% of global revenue under GDPR for compliance violations they believed were resolved > When clients threaten to leave, Delve reportedly pairs them with an external vCISO for manual off-platform work, which the author argues proves their own platform can't deliver real compliance > Delve's sales price dropped from $15,000 to $6,000 with ISO 27001 and a penetration test thrown in when a client mentioned considering a competitor

English
2
1
2
1.2K
Priyaa
Priyaa@pritopian·
@BryanOnel86 Idk enough but when I was evaluating, it felt like it boiled down to pricing. There’s a bunch of stuff you need to get through and diff platforms verbalized it differently. But nothing stood out for a startup like ours as differentiated. May be it changes for larger firms.
English
0
0
0
341
Bryan Onel
Bryan Onel@BryanOnel86·
I would actually argue that there is a ton that can still be done, but they are just really hard. Compliance is mostly about evidence collection, so most platforms are built around automating the retrieval and processing of that data. Building security tools that are part of the compliance platform from day 1 solves a lot of that. If your code scanner, MDM, attack surface management, vulnerability scanning, and more are all part of the same platform, then you don’t need to automate the collection and processing anymore. You can just enforce checks directly through the security tools themselves. That is our play at Oneleet.
English
1
0
12
664
Priyaa
Priyaa@pritopian·
We evaluated many vendors for SOC2. And eventually decided to not move forward with delve. They did seem like nice people, so I hope they come out to clarify. there’s a sea of people selling quick compliance. The harsh truth is it is time consuming. This whole space is commoditized. No amount of AI is accelerating the timeline to get SOC2. 🥲 everyone is just going to have to compete on price or branding.
erin griffith@eringriffith

A detailed and brutal look at the tactics of buzzy AI compliance startup Delve "Delve built a machine designed to make clients complicit without their knowledge, to manufacture plausible deniability while producing exactly the opposite." substack.com/home/post/p-19…

English
3
0
27
6.5K
Priyaa
Priyaa@pritopian·
Priyaa@pritopian

If AI replaces all human labor, and that labor is metered by 3 frontier labs, then we are centralizing cognitive power into a few entities. Back in 2012, TIMES called Amazon an 800 lb. gorilla that you build around, but not compete with. The consensus was they will be the de facto distribution layer. It was true for a while, but what Amazon did was create a new playground and set the rules for competition. (fast deliveries, easy returns, one click purchase etc.) Eventually the merchants realized that Amazon will never let them own the consumer relationship. Worse, Amazon used merchant data to spot what was working and launch competing products. This made Shopify a no brainer to adopt, and led to the DTC movement. The current consensus is AI will largely replace human labor, and that labor will be provisioned by 2 or 3 frontier labs. In that world, companies aren't hiring those AI workers as FTE, they're renting them. And just like with Amazon, it creates concentration risk. Over time, companies will pull back. Open source may not fully catch up but smaller, specialized models built in-house on top of internal private data will become important. They’ll mix and match models across workflows, allocate token spend intentionally, and decide what stays in-house vs. outsourced. It’s funny this @a16z post from 2013 by @jeff_jordan on how to compete with Amazon is the same advice founders get today to stay out of the Eye of Sauron. Sell differentiated product Develop your own products Merchandise product differently Deploy alternative distribution strategies Leverage unique advantages

0
0
0
325
TBPN
TBPN@tbpn·
"There's a foundation model maximalist POV that the labs are going to do everything, in every nook and cranny of the economy." "I have a hard time imagining that version of the future coming to fruition, because people do business with people." - Sequoia partner @gradypb "Between a job to be done, and the raw capabilities of a model, there's a lot that needs to happen to shape it into the path of least resistance for you to travel down as a user to get to the right answer with the least amount of pain." "And there's probably a person in between who's going to do that work. And as a customer, you want to do business with that person."
English
3
10
69
13.7K
Priyaa
Priyaa@pritopian·
@hypersoren Yeah you’ve to acquire or build that capacity faster than the frontier labs’ ability to outpace you to capture proprietary data loop that can unlock the next intelligence leap. We’re seeing this play out with cursor.
English
0
0
1
18
Soren Larson
Soren Larson@hypersoren·
@pritopian wonder how quickly this aggregation / disaggregation dynamic will play out Seems the candidate open disaggregator would need model economics rivaling the labs
English
1
0
1
37
Priyaa
Priyaa@pritopian·
1) Build a model so good that most of them will want to just get started and use it exclusively for idea to production 2) Then slowly folks will want model independence and will move to an aggregator. Hoepfully the aggregator uses the distribution and data to build capacity for a new orchestration layer that treats models like disposable commodities. 3) People get tired, and want to own their intelligence layer + proprietary data loops. Will take stuff in-house with a combo of open source models + fine tuning. New intelligence leap. repeat.
AprilNEA@AprilNEA

7/ This puts Anthropic in direct competition with Vercel, Netlify, Replit, Lovable, and Bolt — but with one advantage nobody else has: they own the entire stack from the LLM to the deployment platform. All discovered via strace, objdump, and go tool objdump on a running Claude Code session. The binary was right there, unstripped, waiting to be read. 🔬

English
1
0
1
442
Priyaa
Priyaa@pritopian·
@ClementDelangue this is exactly what we are enabling for graphic design! decomposing AI visuals into layers, and delegating these layers to specialized models. no point in resetting the design state every single time.
English
1
0
1
142
clem 🤗
clem 🤗@ClementDelangue·
Someone should build a truly multi-model agent that switches between hundreds of different specialized models for different tasks (including even maybe local models ultimately?) Feels like it would increase speed, affordability & powerfulness by an order of magnitude for agents and doable with inference providers on Hugging Face and Hugging Face skills!
English
38
8
123
10.1K
Priyaa retweetledi
kepano
kepano@kepano·
your edge is whatever you know that the models don't know
English
191
460
4.2K
413.6K
Priyaa
Priyaa@pritopian·
side note: that many typos in the query would have sent good old Google search into a frenzy.crazy that now you can dump your stream of thoughts and the LLMs get you (I feel seen. I don’t bother correcting typos or half baked sentences because I know the LLMs would get what I’m saying)
English
0
0
0
72
Sherry Jiang
Sherry Jiang@SherryYanJiang·
playing around with @MagicPathAI and i have to say one of my favorite things about it so far is the ability to use language to control some of the traditional figma functionalities that always felt clunky to me
Sherry Jiang tweet media
English
2
3
9
887